Compare commits

...

19 Commits

Author SHA1 Message Date
baschno
50b0094e86 bump b3sup 2025-12-13 14:25:43 +01:00
Masaki Yatsu
0d45433ea9 fix(litellm): fix langfuse integration 2025-12-11 11:43:37 +09:00
Masaki Yatsu
a8599b66f4 fix(minio): fix OIDC and add public access recipes 2025-12-10 13:26:41 +09:00
Masaki Yatsu
1924e56ad7 chore(ollama): set ollama resource by env-vars 2025-12-10 10:16:05 +09:00
Masaki Yatsu
dae4e9d7ac feat(temporal): enable worker insight 2025-12-08 20:47:19 +09:00
Masaki Yatsu
cfcb278c4d fix(miniflux): fix login error 2025-12-08 13:22:00 +09:00
Masaki Yatsu
022c85c0dc feat(temporal): upgrade Temporal 2025-12-08 09:23:17 +09:00
Masaki Yatsu
3ac8a72df6 docs: write about Temporal 2025-12-07 16:20:17 +09:00
Masaki Yatsu
ca0a8dacba feat(temporal): install Temporal 2025-12-07 16:18:50 +09:00
Masaki Yatsu
fb1e4c20fa chore(ollama): set longer timeout to deploy Ollama 2025-12-06 10:21:52 +09:00
Masaki Yatsu
593da33d64 chore(langfuse): adjust resources 2025-12-05 10:20:16 +09:00
Masaki Yatsu
f3bc41e9eb feat(litellm): litellm -> langfuse integration and some fixes 2025-12-05 10:11:46 +09:00
Masaki Yatsu
98b03704d7 fix(jupyterhub): set sticky sessions for websocket 2025-12-05 10:09:21 +09:00
Masaki Yatsu
6fa0d27f7d feat(nats): install NATS 2025-12-04 15:48:53 +09:00
Masaki Yatsu
d9ee90c32c feat(clickhouse): enable Prometheus monitoring 2025-12-04 11:34:22 +09:00
Masaki Yatsu
7dc732268e docs: write about LiteLLM 2025-12-04 00:21:13 +09:00
Masaki Yatsu
2955d7d783 feat(litellm): SSO and user management 2025-12-04 00:19:14 +09:00
Masaki Yatsu
5055a36d87 feat(litellm): install LiteLLM 2025-12-03 23:05:23 +09:00
Masaki Yatsu
46fdff720f chore(librechat): adjust kubernetes resouces 2025-12-03 21:01:28 +09:00
39 changed files with 3758 additions and 53 deletions

View File

@@ -429,6 +429,7 @@ When writing Markdown documentation:
```
2. **Always validate with markdownlint-cli2**:
- Run `markdownlint-cli2 <file>` before committing any Markdown files
- Run from the project root directory to use `.markdownlint.yaml` config:
`cd <top-dir> && markdownlint-cli2 <relative-path>`
- Fix all linting errors to ensure consistent formatting
- Pay attention to code block language specifications (MD040) and list formatting (MD029)

View File

@@ -63,6 +63,7 @@ A remotely accessible Kubernetes home lab with OIDC authentication. Build a mode
### LLM & AI Applications (Optional)
- **[Ollama](https://ollama.com/)**: Local LLM inference server with GPU acceleration
- **[LiteLLM](https://litellm.ai/)**: Unified LLM gateway for accessing multiple providers through OpenAI-compatible API
- **[LibreChat](https://www.librechat.ai/)**: Web-based chat interface with multi-model support and MCP integration
- **[Langfuse](https://langfuse.com/)**: LLM observability and analytics platform for tracking and debugging AI applications
@@ -70,6 +71,7 @@ A remotely accessible Kubernetes home lab with OIDC authentication. Build a mode
- **[Dagster](https://dagster.io/)**: Modern data orchestration platform
- **[Apache Airflow](https://airflow.apache.org/)**: Workflow orchestration and task scheduling
- **[Temporal](https://temporal.io/)**: Durable workflow execution for distributed applications
### Security & Compliance (Optional)
@@ -346,6 +348,18 @@ LLM observability and analytics platform:
[📖 See Langfuse Documentation](./langfuse/README.md)
### LiteLLM
Unified LLM gateway and proxy:
- **Multi-Provider Support**: Anthropic, OpenAI, Ollama, Mistral, Groq, and more through single API
- **OpenAI-Compatible**: Drop-in replacement for OpenAI SDK
- **Virtual Keys**: Generate scoped API keys for users with usage tracking
- **Cost Tracking**: Monitor spending across all LLM providers
- **Keycloak Authentication**: OAuth2 for Admin UI with role-based access
[📖 See LiteLLM Documentation](./litellm/README.md)
### Dagster
Modern data orchestration platform:
@@ -366,6 +380,17 @@ Workflow orchestration platform:
[📖 See Airflow Documentation](./airflow/README.md)
### Temporal
Durable workflow execution platform:
- **Durable Execution**: Workflows survive process and infrastructure failures
- **Saga Pattern**: Implement distributed transactions with compensating actions
- **Multi-Language SDKs**: Go, Python, TypeScript, Java, .NET, PHP
- **Keycloak Authentication**: OAuth2 for Web UI access
[📖 See Temporal Documentation](./temporal/README.md)
### Fairwinds Polaris
Kubernetes configuration validation and best practices auditing:
@@ -485,6 +510,7 @@ kubectl --context yourpc-oidc get nodes
# JupyterHub: https://jupyter.yourdomain.com
# MLflow: https://mlflow.yourdomain.com
# Langfuse: https://langfuse.yourdomain.com
# LiteLLM: https://litellm.yourdomain.com
# LibreChat: https://chat.yourdomain.com
```

View File

@@ -2,4 +2,5 @@ clickhouse-credentials-external-secret.yaml
clickhouse-ingress.yaml
clickhouse-installation-template.yaml
clickhouse-operator-values.yaml
clickhouse-servicemonitor.yaml
clickhouse.yaml

View File

@@ -44,3 +44,31 @@ ClickHouse can use the following Linux capabilities for enhanced performance, bu
| `SYS_NICE` | Thread priority control via `os_thread_priority` | Setting has no effect |
These capabilities are disabled by default to comply with baseline Pod Security Standards. To enable them, the namespace must allow privileged pods, and you need to uncomment the `add` line in `clickhouse-installation-template.yaml`.
## Monitoring
ClickHouse exposes Prometheus metrics on port 9363. When Prometheus (kube-prometheus-stack) is installed, monitoring can be enabled during installation or manually.
### Enable Monitoring
```bash
just clickhouse::setup-monitoring
```
This creates a ServiceMonitor and a metrics Service for Prometheus to scrape.
### Grafana Dashboard
Import the ClickHouse dashboard from Grafana.com:
1. Open Grafana → **Dashboards****New****Import**
2. Enter Dashboard ID: `14192`
3. Click **Load**, select **Prometheus** data source, then **Import**
The dashboard includes panels for memory, connections, queries, I/O, replication, merge operations, cache, and ZooKeeper metrics.
### Remove Monitoring
```bash
just clickhouse::remove-monitoring
```

View File

@@ -20,6 +20,10 @@ spec:
containers:
- name: clickhouse
image: {{ .Env.CLICKHOUSE_IMAGE }}
ports:
- name: prometheus
containerPort: 9363
protocol: TCP
resources:
requests:
cpu: {{ .Env.CLICKHOUSE_CPU_REQUEST }}

View File

@@ -0,0 +1,67 @@
{{- if .Env.MONITORING_ENABLED }}
---
apiVersion: v1
kind: Service
metadata:
name: clickhouse-metrics
namespace: {{ .Env.CLICKHOUSE_NAMESPACE }}
labels:
app: clickhouse
clickhouse.altinity.com/chi: clickhouse
spec:
type: ClusterIP
ports:
- name: prometheus
port: 9363
targetPort: 9363
protocol: TCP
selector:
clickhouse.altinity.com/chi: clickhouse
---
apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
name: clickhouse
namespace: {{ .Env.CLICKHOUSE_NAMESPACE }}
labels:
app: clickhouse
release: kube-prometheus-stack
spec:
selector:
matchLabels:
app: clickhouse
clickhouse.altinity.com/chi: clickhouse
namespaceSelector:
matchNames:
- {{ .Env.CLICKHOUSE_NAMESPACE }}
endpoints:
- port: prometheus
path: /metrics
interval: 30s
scrapeTimeout: 10s
---
apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
name: clickhouse-operator
namespace: {{ .Env.CLICKHOUSE_NAMESPACE }}
labels:
app: clickhouse-operator
release: kube-prometheus-stack
spec:
selector:
matchLabels:
app.kubernetes.io/name: altinity-clickhouse-operator
namespaceSelector:
matchNames:
- {{ .Env.CLICKHOUSE_NAMESPACE }}
endpoints:
- port: ch-metrics
path: /metrics
interval: 30s
scrapeTimeout: 10s
- port: op-metrics
path: /metrics
interval: 30s
scrapeTimeout: 10s
{{- end }}

View File

@@ -29,6 +29,17 @@ spec:
<schema_type>transposed</schema_type>
</asynchronous_metric_log>
</clickhouse>
# Enable Prometheus metrics endpoint
prometheus.xml: |
<clickhouse>
<prometheus>
<endpoint>/metrics</endpoint>
<port>9363</port>
<metrics>true</metrics>
<events>true</events>
<asynchronous_metrics>true</asynchronous_metrics>
</prometheus>
</clickhouse>
users:
admin/k8s_secret_password: clickhouse-credentials/admin
admin/networks/ip: "::/0"

View File

@@ -5,6 +5,8 @@ export CLICKHOUSE_HOST := env("CLICKHOUSE_HOST", "")
export CLICKHOUSE_CHART_VERSION := env("CLICKHOUSE_CHART_VERSION", "0.25.5")
export CLICKHOUSE_IMAGE := env("CLICKHOUSE_IMAGE", "clickhouse/clickhouse-server:25.10")
export EXTERNAL_SECRETS_NAMESPACE := env("EXTERNAL_SECRETS_NAMESPACE", "external-secrets")
export PROMETHEUS_NAMESPACE := env("PROMETHEUS_NAMESPACE", "monitoring")
export MONITORING_ENABLED := env("MONITORING_ENABLED", "")
# ClickHouse resource settings
export CLICKHOUSE_MEMORY_REQUEST := env("CLICKHOUSE_MEMORY_REQUEST", "1Gi")
@@ -107,6 +109,16 @@ install:
--placeholder="e.g., clickhouse.example.com"
)
done
# Check if Prometheus is available and ask about monitoring
if helm status kube-prometheus-stack -n ${PROMETHEUS_NAMESPACE} &>/dev/null; then
if [ -z "${MONITORING_ENABLED}" ]; then
if gum confirm "Enable Prometheus monitoring?"; then
MONITORING_ENABLED="true"
else
MONITORING_ENABLED="false"
fi
fi
fi
echo "Installing ClickHouse..."
just create-namespace
just install-zookeeper
@@ -124,6 +136,10 @@ install:
kubectl wait --for=jsonpath='{.status.status}'=Completed \
clickhouseinstallation/clickhouse -n ${CLICKHOUSE_NAMESPACE} --timeout=600s
just setup-ingress ${CLICKHOUSE_HOST}
# Setup monitoring if enabled
if [ "${MONITORING_ENABLED}" = "true" ]; then
just setup-monitoring
fi
echo "ClickHouse installation completed successfully"
echo "ClickHouse API at: https://${CLICKHOUSE_HOST}"
@@ -137,6 +153,27 @@ setup-ingress host:
kubectl apply -n ${CLICKHOUSE_NAMESPACE} -f clickhouse-ingress.yaml
echo "ClickHouse Ingress configured successfully"
# Setup Prometheus monitoring for ClickHouse
setup-monitoring:
#!/bin/bash
set -euo pipefail
echo "Setting up Prometheus monitoring for ClickHouse..."
kubectl label namespace ${CLICKHOUSE_NAMESPACE} buun.channel/enable-monitoring=true --overwrite
MONITORING_ENABLED="true" gomplate -f clickhouse-servicemonitor.gomplate.yaml \
-o clickhouse-servicemonitor.yaml
kubectl apply -f clickhouse-servicemonitor.yaml
echo "Prometheus monitoring configured successfully"
# Remove Prometheus monitoring for ClickHouse
remove-monitoring:
#!/bin/bash
set -euo pipefail
echo "Removing Prometheus monitoring for ClickHouse..."
kubectl delete servicemonitor clickhouse clickhouse-operator -n ${CLICKHOUSE_NAMESPACE} --ignore-not-found
kubectl delete service clickhouse-metrics -n ${CLICKHOUSE_NAMESPACE} --ignore-not-found
kubectl label namespace ${CLICKHOUSE_NAMESPACE} buun.channel/enable-monitoring- --ignore-not-found
echo "Prometheus monitoring removed"
# Uninstall ClickHouse (delete_volumes='false' to preserve PVCs and namespace)
uninstall delete-volumes='true':
#!/bin/bash

View File

@@ -33,33 +33,55 @@ install:
)
done
if [ -z "${MINIFLUX_DB_PASSWORD}" ]; then
MINIFLUX_DB_PASSWORD=$(
gum input --prompt="Database password (empty to auto-generate): " \
--width=100 --password
)
# Check if PostgreSQL user already exists
if just postgres::user-exists ${MINIFLUX_DB_USERNAME} &>/dev/null; then
echo "PostgreSQL user '${MINIFLUX_DB_USERNAME}' already exists."
# Use existing password from Vault
if existing_db_pw=$(just vault::get miniflux/db password 2>/dev/null); then
echo "Using existing database password from Vault."
export MINIFLUX_DB_PASSWORD="${existing_db_pw}"
else
echo "Error: User exists but password not found in Vault." >&2
echo "Please delete the user first: just postgres::delete-user ${MINIFLUX_DB_USERNAME}" >&2
exit 1
fi
else
if [ -z "${MINIFLUX_DB_PASSWORD}" ]; then
MINIFLUX_DB_PASSWORD=$(just utils::random-password)
echo "Generated random password: ${MINIFLUX_DB_PASSWORD}"
MINIFLUX_DB_PASSWORD=$(
gum input --prompt="Database password (empty to auto-generate): " \
--width=100 --password
)
if [ -z "${MINIFLUX_DB_PASSWORD}" ]; then
MINIFLUX_DB_PASSWORD=$(just utils::random-password)
echo "Generated random password for database."
fi
fi
export MINIFLUX_DB_PASSWORD
just postgres::create-user-and-db \
${MINIFLUX_DB_USERNAME} ${MINIFLUX_DB_NAME} ${MINIFLUX_DB_PASSWORD}
just vault::put miniflux/db username=${MINIFLUX_DB_USERNAME} \
password=${MINIFLUX_DB_PASSWORD} database=${MINIFLUX_DB_NAME}
fi
just postgres::create-user-and-db \
${MINIFLUX_DB_USERNAME} ${MINIFLUX_DB_NAME} ${MINIFLUX_DB_PASSWORD}
just vault::put miniflux/db username=${MINIFLUX_DB_USERNAME} \
password=${MINIFLUX_DB_PASSWORD} database=${MINIFLUX_DB_NAME}
if [ -z "${MINIFLUX_ADMIN_PASSWORD}" ]; then
MINIFLUX_ADMIN_PASSWORD=$(
gum input --prompt="Admin password (empty to auto-generate): " \
--width=100 --password
)
# Check if admin password exists in Vault
if existing_admin_pw=$(just vault::get miniflux/admin password 2>/dev/null); then
echo "Using existing admin password from Vault."
export MINIFLUX_ADMIN_PASSWORD="${existing_admin_pw}"
else
if [ -z "${MINIFLUX_ADMIN_PASSWORD}" ]; then
MINIFLUX_ADMIN_PASSWORD=$(just utils::random-password)
echo "Generated random password: ${MINIFLUX_ADMIN_PASSWORD}"
MINIFLUX_ADMIN_PASSWORD=$(
gum input --prompt="Admin password (empty to auto-generate): " \
--width=100 --password
)
if [ -z "${MINIFLUX_ADMIN_PASSWORD}" ]; then
MINIFLUX_ADMIN_PASSWORD=$(just utils::random-password)
echo "Generated random password for admin."
fi
fi
export MINIFLUX_ADMIN_PASSWORD
just vault::put miniflux/admin username=${MINIFLUX_ADMIN_USERNAME} \
password=${MINIFLUX_ADMIN_PASSWORD}
fi
just vault::put miniflux/admin username=${MINIFLUX_ADMIN_USERNAME} \
password=${MINIFLUX_ADMIN_PASSWORD}
# https://github.com/gabe565/charts/tree/main/charts/miniflux
MINIFLUX_NAMESPACE=${MINIFLUX_NAMESPACE} \
@@ -80,3 +102,26 @@ admin-username:
# Print admin password
admin-password:
@just vault::get miniflux/admin password
# Reset admin password (deletes admin user from DB so it gets recreated on pod restart)
reset-admin-password:
#!/bin/bash
set -euo pipefail
echo "This will reset the admin password to the value stored in Vault."
if ! gum confirm "Continue?"; then
echo "Cancelled."
exit 0
fi
echo "Deleting admin user from database..."
kubectl exec -n postgres postgres-cluster-1 -c postgres -- \
psql -U postgres -d ${MINIFLUX_DB_NAME} -c "DELETE FROM users WHERE username = '${MINIFLUX_ADMIN_USERNAME}';"
echo "Restarting Miniflux pod..."
kubectl rollout restart deployment/miniflux -n ${MINIFLUX_NAMESPACE}
kubectl rollout status deployment/miniflux -n ${MINIFLUX_NAMESPACE} --timeout=60s
echo ""
echo "Admin password has been reset."
echo "Username: $(just vault::get miniflux/admin username)"
echo "Password: $(just vault::get miniflux/admin password)"

View File

@@ -15,6 +15,8 @@ ingress:
env:
DATABASE_URL: "postgresql://{{ .Env.MINIFLUX_DB_USERNAME }}:{{ .Env.MINIFLUX_DB_PASSWORD }}@postgres-cluster-rw.postgres:5432/{{ .Env.MINIFLUX_DB_NAME }}"
BASE_URL: "https://{{ .Env.MINIFLUX_HOST }}"
HTTPS: "1"
ADMIN_USERNAME: {{ .Env.MINIFLUX_ADMIN_USERNAME }}
ADMIN_PASSWORD: {{ .Env.MINIFLUX_ADMIN_PASSWORD }}

View File

@@ -445,6 +445,10 @@ ingress:
annotations:
kubernetes.io/ingress.class: traefik
traefik.ingress.kubernetes.io/router.entrypoints: websecure
# Enable sticky sessions for WebSocket connections (required for Jupyter RTC/MCP)
traefik.ingress.kubernetes.io/service.sticky.cookie: "true"
traefik.ingress.kubernetes.io/service.sticky.cookie.name: jupyter-session
traefik.ingress.kubernetes.io/service.sticky.cookie.secure: "true"
ingressClassName: traefik
hosts:
- {{ .Env.JUPYTERHUB_HOST }}

View File

@@ -23,10 +23,12 @@ mod kserve
mod langfuse
mod lakekeeper
mod librechat
mod litellm
mod longhorn
mod metabase
mod mlflow
mod minio
mod nats
mod nvidia-device-plugin
mod fairwinds-polaris
mod oauth2-proxy
@@ -37,6 +39,7 @@ mod qdrant
mod querybook
mod security
mod superset
mod temporal
mod trino
mod utils
mod vault

View File

@@ -98,8 +98,8 @@ install:
--version ${LANGFUSE_CHART_VERSION} -n ${LANGFUSE_NAMESPACE} --wait \
-f langfuse-values.yaml
# Uninstall Langfuse
uninstall:
# Uninstall Langfuse (delete-data: true to delete database and storage)
uninstall delete-data='false':
#!/bin/bash
set -euo pipefail
helm uninstall langfuse -n ${LANGFUSE_NAMESPACE} --wait --ignore-not-found
@@ -108,19 +108,29 @@ uninstall:
# Clean up Keycloak client and Vault secrets to avoid stale credentials
just delete-keycloak-client || true
echo "Langfuse uninstalled successfully"
echo ""
echo "Note: The following resources were NOT deleted:"
echo " - PostgreSQL user and database (langfuse)"
echo " - ClickHouse user and database (langfuse)"
echo " - MinIO user and bucket (langfuse)"
echo " - Keycloak user (langfuse)"
echo ""
echo "To delete these resources, run:"
echo " just langfuse::delete-postgres-user-and-db"
echo " just langfuse::delete-clickhouse-user"
echo " just langfuse::delete-minio-user"
echo " just langfuse::delete-keycloak-user"
if [ "{{ delete-data }}" = "true" ]; then
echo "Deleting database and storage..."
just delete-postgres-user-and-db || true
just delete-clickhouse-user || true
just delete-minio-user || true
just delete-keycloak-user || true
just delete-salt || true
just delete-nextauth-secret || true
just delete-redis-password || true
echo "Langfuse uninstalled with all data deleted."
else
echo "Langfuse uninstalled successfully"
echo ""
echo "Note: The following resources were NOT deleted:"
echo " - PostgreSQL user and database (langfuse)"
echo " - ClickHouse user and database (langfuse)"
echo " - MinIO user and bucket (langfuse)"
echo " - Keycloak user (langfuse)"
echo " - Vault secrets (langfuse/*)"
echo ""
echo "To delete all data, run:"
echo " just langfuse::uninstall true"
fi
# Create all secrets (PostgreSQL, Keycloak, MinIO, Redis)
create-secrets:

View File

@@ -69,13 +69,13 @@ langfuse:
tls:
enabled: true
# Resource configuration based on Goldilocks/VPA recommendations
# CPU limits increased to handle startup spikes
# Resource recommendations from Goldilocks VPA
# web target: cpu=15m, memory=717Mi
web:
resources:
requests:
cpu: 15m
memory: 704Mi
cpu: 25m
memory: 768Mi
limits:
cpu: 100m
memory: 1.5Gi
@@ -89,10 +89,12 @@ langfuse:
timeoutSeconds: 30
failureThreshold: 5
# Resource recommendations from Goldilocks VPA
# worker target: cpu=15m, memory=380Mi
worker:
resources:
requests:
cpu: 15m
cpu: 25m
memory: 512Mi
limits:
cpu: 100m
@@ -113,6 +115,16 @@ redis:
username: "default"
existingSecret: redis-auth
existingSecretPasswordKey: secret
# Resource recommendations from Goldilocks VPA
# valkey target: cpu=15m, memory=100Mi
master:
resources:
requests:
cpu: 25m
memory: 128Mi
limits:
cpu: 100m
memory: 256Mi
clickhouse:
deploy: false

View File

@@ -94,10 +94,10 @@ ingress:
resources:
requests:
cpu: 100m
cpu: 25m
memory: 512Mi
limits:
cpu: 1000m
cpu: 100m
memory: 1Gi
mongodb:
@@ -110,6 +110,13 @@ mongodb:
tag: "latest"
persistence:
size: 8Gi
resources:
requests:
cpu: 75m
memory: 512Mi
limits:
cpu: 500m
memory: 2Gi
podSecurityContext:
fsGroup: 1001
seccompProfile:
@@ -132,6 +139,13 @@ meilisearch:
tag: "v1.7.3"
auth:
existingMasterKeySecret: "librechat-credentials-env"
resources:
requests:
cpu: 25m
memory: 256Mi
limits:
cpu: 100m
memory: 1Gi
podSecurityContext:
fsGroup: 1000
seccompProfile:

3
litellm/.gitignore vendored Normal file
View File

@@ -0,0 +1,3 @@
litellm-values.yaml
apikey-external-secret.yaml
models.yaml

547
litellm/README.md Normal file
View File

@@ -0,0 +1,547 @@
# LiteLLM
Unified LLM gateway and proxy for accessing multiple LLM providers through a single OpenAI-compatible API:
- **Multi-Provider Support**: Anthropic, OpenAI, Ollama, Mistral, Groq, Cohere, Azure, Bedrock, Vertex AI
- **OpenAI-Compatible API**: Drop-in replacement for OpenAI SDK
- **Load Balancing & Fallback**: Automatic failover between providers
- **Virtual Keys**: Generate API keys for users with usage tracking
- **Cost Tracking**: Monitor spending across providers
- **Rate Limiting**: Control usage per key/user
## Prerequisites
- Kubernetes cluster (k3s)
- External Secrets Operator (required)
- PostgreSQL cluster (CloudNativePG)
- Vault for secrets management
## Configuration Overview
LiteLLM requires two types of configuration:
1. **Environment variables** (`.env.local`): Host, namespace, chart version
2. **Model definitions** (`models.yaml`): LLM providers and models to expose
This separation allows flexible model configuration without modifying environment files.
## Installation
### Step 1: Create Model Configuration
Copy the example configuration and customize:
```bash
cp litellm/models.example.yaml litellm/models.yaml
```
Edit `litellm/models.yaml` to configure your models:
```yaml
# Anthropic Claude
- model_name: claude-sonnet
litellm_params:
model: anthropic/claude-3-7-sonnet-latest
api_key: os.environ/ANTHROPIC_API_KEY
# OpenAI
- model_name: gpt-4o
litellm_params:
model: openai/gpt-4o
api_key: os.environ/OPENAI_API_KEY
# Ollama (local models - no API key required)
- model_name: llama3
litellm_params:
model: ollama/llama3.2
api_base: http://ollama.ollama:11434
```
### Step 2: Set API Keys
For each provider that requires an API key:
```bash
just litellm::set-api-key anthropic
just litellm::set-api-key openai
```
Or interactively select the provider:
```bash
just litellm::set-api-key
```
API keys are stored in Vault and synced to Kubernetes via External Secrets Operator.
### Step 3: Install LiteLLM
```bash
just litellm::install
```
You will be prompted for:
- **LiteLLM host (FQDN)**: e.g., `litellm.example.com`
- **Enable Prometheus monitoring**: If kube-prometheus-stack is installed
## Model Management
### Add a Model Interactively
```bash
just litellm::add-model
```
This guides you through:
1. Selecting a provider
2. Choosing a model
3. Setting a model alias
### Remove a Model
```bash
just litellm::remove-model
```
### List Configured Models
```bash
just litellm::list-models
```
### Example Output
```text
Configured models:
- claude-sonnet: anthropic/claude-3-7-sonnet-latest
- claude-haiku: anthropic/claude-3-5-haiku-latest
- llama3: ollama/llama3.2
```
## API Key Management
### Set API Key for a Provider
```bash
just litellm::set-api-key anthropic
```
### Get API Key (from Vault)
```bash
just litellm::get-api-key anthropic
```
### Verify All Required Keys
```bash
just litellm::verify-api-keys
```
## Environment Variables
| Variable | Default | Description |
| -------- | ------- | ----------- |
| `LITELLM_NAMESPACE` | `litellm` | Kubernetes namespace |
| `LITELLM_CHART_VERSION` | `0.1.825` | Helm chart version |
| `LITELLM_HOST` | (prompt) | External hostname (FQDN) |
| `OLLAMA_NAMESPACE` | `ollama` | Ollama namespace for local models |
| `MONITORING_ENABLED` | (prompt) | Enable Prometheus ServiceMonitor |
## Authentication
LiteLLM has two types of authentication:
1. **API Access**: Uses Master Key or Virtual Keys for programmatic access
2. **Admin UI**: Uses Keycloak SSO for browser-based access
### Enable SSO for Admin UI
After installing LiteLLM, enable Keycloak authentication for the Admin UI:
```bash
just litellm::setup-oidc
```
This will:
- Create a Keycloak client for LiteLLM
- Store the client secret in Vault
- Configure LiteLLM with OIDC environment variables
- Upgrade the deployment with SSO enabled
### Disable SSO
To disable SSO and return to unauthenticated Admin UI access:
```bash
just litellm::disable-oidc
```
### SSO Configuration Details
| Setting | Value |
| ------- | ----- |
| Callback URL | `https://<litellm-host>/sso/callback` |
| Authorization Endpoint | `https://<keycloak-host>/realms/<realm>/protocol/openid-connect/auth` |
| Token Endpoint | `https://<keycloak-host>/realms/<realm>/protocol/openid-connect/token` |
| Userinfo Endpoint | `https://<keycloak-host>/realms/<realm>/protocol/openid-connect/userinfo` |
| Scope | `openid email profile` |
## User Management
SSO users are automatically created in LiteLLM when they first log in. By default, new users are assigned the `internal_user_viewer` role (read-only access).
### List Users
```bash
just litellm::list-users
```
### Assign Role to User
Interactively select user and role:
```bash
just litellm::assign-role
```
Or specify directly:
```bash
just litellm::assign-role buun proxy_admin
```
### User Roles
| Role | Description |
| ---- | ----------- |
| `proxy_admin` | Full admin access (manage keys, users, models, settings) |
| `proxy_admin_viewer` | Admin read-only access |
| `internal_user` | Can create and manage own API keys |
| `internal_user_viewer` | Read-only access (default for SSO users) |
**Note**: To manage API keys in the Admin UI, users need at least `internal_user` or `proxy_admin` role.
## API Usage
LiteLLM exposes an OpenAI-compatible API at `https://your-litellm-host/`.
### Get Master Key
```bash
just litellm::master-key
```
### Generate Virtual Key for a User
```bash
just litellm::generate-virtual-key buun
```
This will prompt for a model selection and generate an API key for the specified user. Select `all` to grant access to all models.
### OpenAI SDK Example
```python
from openai import OpenAI
client = OpenAI(
base_url="https://litellm.example.com",
api_key="sk-..." # Virtual key or master key
)
response = client.chat.completions.create(
model="claude-sonnet", # Use your model alias
messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)
```
### curl Example
```bash
curl https://litellm.example.com/v1/chat/completions \
-H "Authorization: Bearer sk-..." \
-H "Content-Type: application/json" \
-d '{
"model": "claude-sonnet",
"messages": [{"role": "user", "content": "Hello!"}]
}'
```
## Team Management
Teams allow you to group users and configure team-specific settings such as Langfuse projects for observability.
### Create a Team
```bash
just litellm::create-team
```
Or with a name directly:
```bash
just litellm::create-team name="project-alpha"
```
### List Teams
```bash
just litellm::list-teams
```
### Get Team Info
```bash
just litellm::get-team team_id=<team-id>
```
### Delete a Team
```bash
just litellm::delete-team team_id=<team-id>
```
### Generate Virtual Key for a Team
```bash
just litellm::generate-team-key
```
This will prompt for team selection and username. The generated key inherits the team's settings (including Langfuse project configuration).
## Langfuse Integration
[Langfuse](https://langfuse.com/) provides LLM observability with tracing, monitoring, and analytics. LiteLLM can send traces to Langfuse for every API call.
### Enable Langfuse Integration
During installation (`just litellm::install`) or upgrade (`just litellm::upgrade`), you will be prompted to enable Langfuse integration. Alternatively:
```bash
just litellm::setup-langfuse
```
You will need Langfuse API keys (Public Key and Secret Key) from the Langfuse UI: **Settings > API Keys**.
### Set Langfuse API Keys
```bash
just litellm::set-langfuse-keys
```
### Disable Langfuse Integration
```bash
just litellm::disable-langfuse
```
### Per-Team Langfuse Projects
Each team can have its own Langfuse project for isolated observability. This is useful when different projects or departments need separate trace data.
#### Setup Flow
1. Create a team:
```bash
just litellm::create-team name="project-alpha"
```
2. Create a Langfuse project for the team and get API keys from Langfuse UI
3. Configure the team's Langfuse project:
```bash
just litellm::set-team-langfuse-project
```
This will prompt for team selection and Langfuse API keys.
4. Generate a key for the team:
```bash
just litellm::generate-team-key
```
5. Use the team key for API calls - traces will be sent to the team's Langfuse project
#### Architecture
```plain
LiteLLM Proxy
|
+-- Default Langfuse Project (for keys without team)
|
+-- Team A --> Langfuse Project A
|
+-- Team B --> Langfuse Project B
```
### Environment Variables
| Variable | Default | Description |
| -------- | ------- | ----------- |
| `LITELLM_LANGFUSE_INTEGRATION_ENABLED` | (prompt) | Enable Langfuse integration |
| `LANGFUSE_HOST` | (prompt) | Langfuse instance hostname |
## Supported Providers
| Provider | Model Prefix | API Key Required |
| -------- | ------------ | ---------------- |
| Anthropic | `anthropic/` | Yes |
| OpenAI | `openai/` | Yes |
| Ollama | `ollama/` | No (uses `api_base`) |
| Mistral | `mistral/` | Yes |
| Groq | `groq/` | Yes |
| Cohere | `cohere/` | Yes |
| Azure OpenAI | `azure/` | Yes |
| AWS Bedrock | `bedrock/` | Yes |
| Google Vertex AI | `vertexai/` | Yes |
## Architecture
```plain
External Users/Applications
|
Cloudflare Tunnel (HTTPS)
|
Traefik Ingress (HTTPS)
|
LiteLLM Proxy (HTTP inside cluster)
|-- PostgreSQL (usage tracking, virtual keys)
|-- Redis (caching, rate limiting)
|-- External Secrets (API keys from Vault)
|
+-- Anthropic API
+-- OpenAI API
+-- Ollama (local)
+-- Other providers...
```
## Upgrade
After modifying `models.yaml` or updating API keys:
```bash
just litellm::upgrade
```
## Uninstall
```bash
just litellm::uninstall
```
This removes:
- Helm release and all Kubernetes resources
- Namespace
- External Secrets
**Note**: The following resources are NOT deleted:
- PostgreSQL database (use `just postgres::delete-db litellm`)
- API keys in Vault
### Full Cleanup
To remove everything including database and Vault secrets:
```bash
just litellm::cleanup
```
## Troubleshooting
### Check Pod Status
```bash
kubectl get pods -n litellm
```
Expected pods:
- `litellm-*` - LiteLLM proxy
- `litellm-redis-master-0` - Redis instance
### View Logs
```bash
kubectl logs -n litellm deployment/litellm --tail=100
```
### API Key Not Working
Verify the ExternalSecret is synced:
```bash
kubectl get externalsecret -n litellm
kubectl get secret apikey -n litellm -o yaml
```
### Model Not Found
Ensure the model is configured in `models.yaml` and the deployment is updated:
```bash
just litellm::list-models
just litellm::upgrade
```
### Provider API Errors
Check if the API key is set correctly:
```bash
just litellm::get-api-key anthropic
```
If empty, set the API key:
```bash
just litellm::set-api-key anthropic
```
### Database Connection Issues
Check PostgreSQL connectivity:
```bash
kubectl exec -n litellm deployment/litellm -- \
psql -h postgres-cluster-rw.postgres -U litellm -d litellm -c "SELECT 1"
```
## Configuration Files
| File | Description |
| ---- | ----------- |
| `models.yaml` | Model definitions (user-created, gitignored) |
| `models.example.yaml` | Example model configuration |
| `litellm-values.gomplate.yaml` | Helm values template |
| `apikey-external-secret.gomplate.yaml` | ExternalSecret for API keys |
| `keycloak-auth-external-secret.gomplate.yaml` | ExternalSecret for Keycloak OIDC |
| `langfuse-auth-external-secret.gomplate.yaml` | ExternalSecret for Langfuse API keys |
## Security Considerations
- **Pod Security Standards**: Namespace configured with **baseline** enforcement
(LiteLLM's Prisma requires write access to `/.cache`, which prevents `restricted` level)
- **Secrets Management**: API keys stored in Vault, synced via External Secrets Operator
- **Virtual Keys**: Generate scoped API keys for users instead of sharing master key
- **TLS/HTTPS**: All external traffic encrypted via Traefik Ingress
- **Database Credentials**: Unique PostgreSQL user with minimal privileges
## References
- [LiteLLM Documentation](https://docs.litellm.ai/)
- [LiteLLM GitHub](https://github.com/BerriAI/litellm)
- [LiteLLM Helm Chart](https://github.com/BerriAI/litellm/tree/main/deploy/charts/litellm-helm)
- [Supported Models](https://docs.litellm.ai/docs/providers)
- [Virtual Keys](https://docs.litellm.ai/docs/proxy/virtual_keys)
- [Langfuse Integration](https://docs.litellm.ai/docs/proxy/logging#langfuse)
- [Team-based Logging](https://docs.litellm.ai/docs/proxy/team_logging)

View File

@@ -0,0 +1,29 @@
{{- $models := (datasource "models") -}}
{{- $providerMap := dict -}}
{{- range $models -}}
{{- if has .litellm_params "api_key" -}}
{{- $parts := strings.Split "/" .litellm_params.model -}}
{{- $provider := index $parts 0 -}}
{{- $providerMap = merge $providerMap (dict $provider true) -}}
{{- end -}}
{{- end -}}
apiVersion: external-secrets.io/v1
kind: ExternalSecret
metadata:
name: apikey-external-secret
namespace: {{ .Env.LITELLM_NAMESPACE }}
spec:
refreshInterval: 1h
secretStoreRef:
name: vault-secret-store
kind: ClusterSecretStore
target:
name: apikey
creationPolicy: Owner
data:
{{- range $provider, $_ := $providerMap }}
- secretKey: {{ $provider | strings.ToUpper }}_API_KEY
remoteRef:
key: litellm/{{ $provider }}
property: apikey
{{- end }}

1074
litellm/justfile Normal file

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,22 @@
apiVersion: external-secrets.io/v1
kind: ExternalSecret
metadata:
name: keycloak-auth-external-secret
namespace: {{ .Env.LITELLM_NAMESPACE }}
spec:
refreshInterval: 1h
secretStoreRef:
name: vault-secret-store
kind: ClusterSecretStore
target:
name: keycloak-auth
creationPolicy: Owner
data:
- secretKey: GENERIC_CLIENT_ID
remoteRef:
key: keycloak/client/litellm
property: client_id
- secretKey: GENERIC_CLIENT_SECRET
remoteRef:
key: keycloak/client/litellm
property: client_secret

View File

@@ -0,0 +1,22 @@
apiVersion: external-secrets.io/v1
kind: ExternalSecret
metadata:
name: langfuse-auth-external-secret
namespace: {{ .Env.LITELLM_NAMESPACE }}
spec:
refreshInterval: 1h
secretStoreRef:
name: vault-secret-store
kind: ClusterSecretStore
target:
name: langfuse-auth
creationPolicy: Owner
data:
- secretKey: LANGFUSE_PUBLIC_KEY
remoteRef:
key: litellm/langfuse
property: public_key
- secretKey: LANGFUSE_SECRET_KEY
remoteRef:
key: litellm/langfuse
property: secret_key

View File

@@ -0,0 +1,114 @@
# https://github.com/BerriAI/litellm/tree/main/deploy/charts/litellm-helm
# https://github.com/BerriAI/litellm/tree/main/litellm/proxy/example_config_yaml
masterkeySecretName: ""
masterkeySecretKey: ""
# Note: LiteLLM image requires write access to /.cache for Prisma
# Pod Security Standards must be set to "baseline" for this namespace
podSecurityContext: {}
securityContext: {}
# Resource recommendations from Goldilocks VPA
# litellm target: cpu=11m, memory=549Mi
resources:
requests:
cpu: 25m
memory: 512Mi
limits:
cpu: 100m
memory: 1Gi
migrationJob:
resources:
requests:
cpu: 100m
memory: 512Mi
limits:
memory: 1Gi
environmentSecrets:
- apikey
{{- if .Env.LITELLM_OIDC_ENABLED }}
- keycloak-auth
{{- end }}
{{- if .Env.LITELLM_LANGFUSE_INTEGRATION_ENABLED }}
- langfuse-auth
{{- end }}
extraEnvVars:
{{- if .Env.LITELLM_OIDC_ENABLED }}
- name: PROXY_BASE_URL
value: "https://{{ .Env.LITELLM_HOST }}"
- name: GENERIC_AUTHORIZATION_ENDPOINT
value: "https://{{ .Env.KEYCLOAK_HOST }}/realms/{{ .Env.KEYCLOAK_REALM }}/protocol/openid-connect/auth"
- name: GENERIC_TOKEN_ENDPOINT
value: "https://{{ .Env.KEYCLOAK_HOST }}/realms/{{ .Env.KEYCLOAK_REALM }}/protocol/openid-connect/token"
- name: GENERIC_USERINFO_ENDPOINT
value: "https://{{ .Env.KEYCLOAK_HOST }}/realms/{{ .Env.KEYCLOAK_REALM }}/protocol/openid-connect/userinfo"
- name: GENERIC_SCOPE
value: "openid email profile"
{{- end }}
{{- if .Env.LITELLM_LANGFUSE_INTEGRATION_ENABLED }}
- name: LANGFUSE_HOST
value: "https://{{ .Env.LANGFUSE_HOST }}"
{{- end }}
proxy_config:
model_list:
{{ file.Read "models.yaml" | indent 4 }}
{{- if .Env.LITELLM_LANGFUSE_INTEGRATION_ENABLED }}
litellm_settings:
success_callback: ["langfuse"]
failure_callback: ["langfuse"]
{{- end }}
db:
useExisting: true
endpoint: postgres-cluster-rw.postgres
database: litellm
secret:
name: postgres-auth
usernameKey: username
passwordKey: password
deployStandalone: false
redis:
enabled: true
# Resource recommendations from Goldilocks VPA
# redis target: cpu=15m, memory=100Mi
master:
resources:
requests:
cpu: 25m
memory: 128Mi
limits:
cpu: 100m
memory: 256Mi
ingress:
enabled: true
className: traefik
annotations:
kubernetes.io/ingress.class: traefik
traefik.ingress.kubernetes.io/router.entrypoints: websecure
hosts:
- host: {{ .Env.LITELLM_HOST }}
paths:
- path: /
pathType: ImplementationSpecific
tls:
- hosts:
- {{ .Env.LITELLM_HOST }}
{{- if .Env.MONITORING_ENABLED }}
serviceMonitor:
enabled: true
labels:
release: kube-prometheus-stack
interval: 30s
scrapeTimeout: 10s
{{- end }}

106
litellm/models.example.yaml Normal file
View File

@@ -0,0 +1,106 @@
# LiteLLM Model Configuration
# Copy this file to models.yaml and customize for your environment.
#
# Usage:
# cp litellm/models.example.yaml litellm/models.yaml
# # Edit models.yaml to add/remove models
# just litellm::install
#
# API keys are stored in Vault and injected as environment variables.
# Use: just litellm::set-api-key provider=<provider>
#
# Supported providers:
# - anthropic: Claude models (Opus, Sonnet, Haiku)
# - openai: GPT and o-series models
# - ollama: Local models (no API key required)
# - azure: Azure OpenAI
# - bedrock: AWS Bedrock
# - vertexai: Google Vertex AI
# - mistral: Mistral AI
# - groq: Groq (fast inference)
# - cohere: Cohere
# Anthropic Claude (https://docs.anthropic.com/en/docs/about-claude/models/overview)
- model_name: claude-sonnet
litellm_params:
model: anthropic/claude-sonnet-4-20250514
api_key: os.environ/ANTHROPIC_API_KEY
- model_name: claude-haiku
litellm_params:
model: anthropic/claude-haiku-4-20251015
api_key: os.environ/ANTHROPIC_API_KEY
# - model_name: claude-opus
# litellm_params:
# model: anthropic/claude-opus-4-20250514
# api_key: os.environ/ANTHROPIC_API_KEY
# OpenAI (https://platform.openai.com/docs/models)
# - model_name: gpt-4o
# litellm_params:
# model: openai/gpt-4o
# api_key: os.environ/OPENAI_API_KEY
# - model_name: gpt-4o-mini
# litellm_params:
# model: openai/gpt-4o-mini
# api_key: os.environ/OPENAI_API_KEY
# - model_name: o3
# litellm_params:
# model: openai/o3
# api_key: os.environ/OPENAI_API_KEY
# - model_name: o4-mini
# litellm_params:
# model: openai/o4-mini
# api_key: os.environ/OPENAI_API_KEY
# Ollama (local models - no API key required)
# - model_name: llama4-scout
# litellm_params:
# model: ollama/llama4:scout
# api_base: http://ollama.ollama:11434
# - model_name: qwen3
# litellm_params:
# model: ollama/qwen3:8b
# api_base: http://ollama.ollama:11434
# - model_name: deepseek-r1
# litellm_params:
# model: ollama/deepseek-r1:8b
# api_base: http://ollama.ollama:11434
# Mistral AI (https://docs.mistral.ai/getting-started/models/models_overview/)
# - model_name: mistral-large
# litellm_params:
# model: mistral/mistral-large-latest
# api_key: os.environ/MISTRAL_API_KEY
# - model_name: ministral-8b
# litellm_params:
# model: mistral/ministral-8b-latest
# api_key: os.environ/MISTRAL_API_KEY
# - model_name: codestral
# litellm_params:
# model: mistral/codestral-latest
# api_key: os.environ/MISTRAL_API_KEY
# Groq (fast inference - https://console.groq.com/docs/models)
# - model_name: groq-llama4-scout
# litellm_params:
# model: groq/meta-llama/llama-4-scout-17b-16e-instruct
# api_key: os.environ/GROQ_API_KEY
# - model_name: groq-llama3.3
# litellm_params:
# model: groq/llama-3.3-70b-versatile
# api_key: os.environ/GROQ_API_KEY
# - model_name: groq-llama3.1
# litellm_params:
# model: groq/llama-3.1-8b-instant
# api_key: os.environ/GROQ_API_KEY

View File

@@ -221,6 +221,83 @@ just minio::bucket-exists mybucket
This returns exit code 0 if the bucket exists, 1 otherwise.
## Public Access
MinIO allows you to configure anonymous (public) access to buckets or specific prefixes for serving static content like images.
### Set Public Download Access
Enable public read access for a bucket or prefix:
```bash
# Set public access for entire bucket
just minio::set-public-download mybucket
# Set public access for specific prefix only
just minio::set-public-download mybucket/public
```
After setting public access, files can be accessed without authentication:
```text
https://your-minio-host/mybucket/public/image.png
```
### Check Public Access Status
View current anonymous access policy:
```bash
just minio::show-public-access mybucket
```
Possible values:
- `private`: No anonymous access (default)
- `download`: Public read access
- `upload`: Public write access
- `public`: Public read and write access
- `custom`: Custom policy applied
### Remove Public Access
Revoke anonymous access:
```bash
just minio::remove-public-access mybucket/public
```
### Using mc Commands
```bash
# Set public download (read-only)
mc anonymous set download myminio/mybucket/public
# Set public upload (write-only)
mc anonymous set upload myminio/mybucket/uploads
# Set full public access (read and write)
mc anonymous set public myminio/mybucket
# Remove public access
mc anonymous set none myminio/mybucket
# Check current policy
mc anonymous get myminio/mybucket
```
### Presigned URLs (Temporary Access)
For temporary access to private objects, use presigned URLs:
```bash
# Generate URL valid for 7 days
mc share download myminio/mybucket/private-file.pdf --expire=168h
# Generate upload URL
mc share upload myminio/mybucket/uploads/ --expire=1h
```
## User Management
### Create MinIO User

View File

@@ -47,6 +47,7 @@ create-root-credentials:
gomplate -f minio-root-external-secret.gomplate.yaml | kubectl apply -f -
echo "Waiting for ExternalSecret to sync..."
sleep 2
kubectl wait --for=condition=Ready externalsecret/minio \
-n ${MINIO_NAMESPACE} --timeout=60s
else
@@ -96,7 +97,12 @@ install:
--placeholder="e.g., minio-console.example.com"
)
fi
# Generate OIDC client secret for confidential client
OIDC_CLIENT_SECRET=$(just utils::random-password)
just keycloak::create-client realm=${KEYCLOAK_REALM} client_id=${MINIO_OIDC_CLIENT_ID} \
client_secret="${OIDC_CLIENT_SECRET}" \
redirect_url="https://${MINIO_HOST}/oauth_callback,https://${MINIO_CONSOLE_HOST}/oauth_callback"
just add-keycloak-minio-policy
just create-namespace
@@ -105,6 +111,28 @@ install:
pod-security.kubernetes.io/enforce=restricted --overwrite
just create-root-credentials
# Store OIDC client secret
if helm status external-secrets -n ${EXTERNAL_SECRETS_NAMESPACE} &>/dev/null; then
echo "Storing OIDC client secret in Vault..."
just vault::put minio/oidc client_id="${MINIO_OIDC_CLIENT_ID}" client_secret="${OIDC_CLIENT_SECRET}"
kubectl delete externalsecret minio-oidc -n ${MINIO_NAMESPACE} --ignore-not-found
gomplate -f minio-oidc-external-secret.gomplate.yaml | kubectl apply -f -
echo "Waiting for ExternalSecret to sync..."
sleep 2
kubectl wait --for=condition=Ready externalsecret/minio-oidc \
-n ${MINIO_NAMESPACE} --timeout=60s
else
echo "Creating OIDC client secret directly..."
kubectl delete secret minio-oidc -n ${MINIO_NAMESPACE} --ignore-not-found
kubectl create secret generic minio-oidc -n ${MINIO_NAMESPACE} \
--from-literal=clientId="${MINIO_OIDC_CLIENT_ID}" \
--from-literal=clientSecret="${OIDC_CLIENT_SECRET}"
if helm status vault -n ${K8S_VAULT_NAMESPACE} &>/dev/null; then
just vault::put minio/oidc client_id="${MINIO_OIDC_CLIENT_ID}" client_secret="${OIDC_CLIENT_SECRET}"
fi
fi
just add-helm-repo
gomplate -f minio-values.gomplate.yaml -o minio-values.yaml
helm upgrade --install minio minio/minio \
@@ -260,3 +288,70 @@ grant-policy user='' policy='':
mc admin policy attach local ${POLICY} --user=${USER}"
echo "✅ Policy ${POLICY} granted to user ${USER}"
# Set public download access for a bucket or prefix
set-public-download path='':
#!/bin/bash
set -euo pipefail
PATH_ARG="{{ path }}"
while [ -z "${PATH_ARG}" ]; do
PATH_ARG=$(
gum input --prompt="Bucket/prefix path: " --width=100 \
--placeholder="e.g., my-bucket/public"
)
done
ROOT_USER=$(just root-username)
ROOT_PASSWORD=$(just root-password)
kubectl -n ${MINIO_NAMESPACE} exec deploy/minio -- \
mc alias set local http://localhost:9000 ${ROOT_USER} ${ROOT_PASSWORD}
kubectl -n ${MINIO_NAMESPACE} exec deploy/minio -- \
mc anonymous set download local/${PATH_ARG}
echo "✅ Public download access enabled for ${PATH_ARG}"
# Remove public access from a bucket or prefix
remove-public-access path='':
#!/bin/bash
set -euo pipefail
PATH_ARG="{{ path }}"
while [ -z "${PATH_ARG}" ]; do
PATH_ARG=$(
gum input --prompt="Bucket/prefix path: " --width=100 \
--placeholder="e.g., my-bucket/public"
)
done
ROOT_USER=$(just root-username)
ROOT_PASSWORD=$(just root-password)
kubectl -n ${MINIO_NAMESPACE} exec deploy/minio -- \
mc alias set local http://localhost:9000 ${ROOT_USER} ${ROOT_PASSWORD}
kubectl -n ${MINIO_NAMESPACE} exec deploy/minio -- \
mc anonymous set none local/${PATH_ARG}
echo "✅ Public access removed from ${PATH_ARG}"
# Show anonymous access policy for a bucket or prefix
show-public-access path='':
#!/bin/bash
set -euo pipefail
PATH_ARG="{{ path }}"
while [ -z "${PATH_ARG}" ]; do
PATH_ARG=$(
gum input --prompt="Bucket/prefix path: " --width=100 \
--placeholder="e.g., my-bucket"
)
done
ROOT_USER=$(just root-username)
ROOT_PASSWORD=$(just root-password)
kubectl -n ${MINIO_NAMESPACE} exec deploy/minio -- \
mc alias set local http://localhost:9000 ${ROOT_USER} ${ROOT_PASSWORD}
kubectl -n ${MINIO_NAMESPACE} exec deploy/minio -- \
mc anonymous get local/${PATH_ARG}

View File

@@ -0,0 +1,22 @@
apiVersion: external-secrets.io/v1
kind: ExternalSecret
metadata:
name: minio-oidc
namespace: {{ .Env.MINIO_NAMESPACE }}
spec:
refreshInterval: 1h
secretStoreRef:
name: vault-secret-store
kind: ClusterSecretStore
target:
name: minio-oidc
creationPolicy: Owner
data:
- secretKey: clientId
remoteRef:
key: minio/oidc
property: client_id
- secretKey: clientSecret
remoteRef:
key: minio/oidc
property: client_secret

View File

@@ -7,8 +7,9 @@ existingSecret: "minio"
oidc:
enabled: true
configUrl: "https://{{ .Env.KEYCLOAK_HOST }}/realms/{{ .Env.KEYCLOAK_REALM }}/.well-known/openid-configuration"
clientId: "{{ .Env.MINIO_OIDC_CLIENT_ID }}"
clientSecret: ""
existingClientSecretName: "minio-oidc"
existingClientIdKey: "clientId"
existingClientSecretKey: "clientSecret"
claimName: "minioPolicy"
scopes: "openid,profile,email"
redirectUri: "https://{{ .Env.MINIO_CONSOLE_HOST }}/oauth_callback"

View File

@@ -3,11 +3,11 @@ gomplate = "4.3.3"
gum = "0.16.2"
helm = "3.19.0"
just = "1.42.4"
k3sup = "0.13.10"
k3sup = "0.13.11"
kubelogin = "1.34.0"
node = "22.18.0"
python = "3.12.11"
telepresence = "2.25.0"
telepresence = "2.25.1"
trivy = "0.67.2"
uv = "0.8.7"
vault = "1.20.2"

113
nats/justfile Normal file
View File

@@ -0,0 +1,113 @@
set fallback := true
export NATS_NAMESPACE := env("NATS_NAMESPACE", "nats")
export NATS_CHART_VERSION := env("NATS_CHART_VERSION", "2.12.2")
export NATS_REPLICAS := env("NATS_REPLICAS", "1")
export NATS_JETSTREAM_ENABLED := env("NATS_JETSTREAM_ENABLED", "true")
export NATS_JETSTREAM_STORAGE_SIZE := env("NATS_JETSTREAM_STORAGE_SIZE", "10Gi")
export NATS_JETSTREAM_MEMORY_SIZE := env("NATS_JETSTREAM_MEMORY_SIZE", "256Mi")
[private]
default:
@just --list --unsorted --list-submodules
# Add Helm repository
add-helm-repo:
helm repo add nats https://nats-io.github.io/k8s/helm/charts/
helm repo update nats
# Remove Helm repository
remove-helm-repo:
helm repo remove nats
# Create NATS namespace
create-namespace:
#!/bin/bash
set -euo pipefail
if ! kubectl get namespace ${NATS_NAMESPACE} &>/dev/null; then
kubectl create namespace ${NATS_NAMESPACE}
fi
kubectl label namespace ${NATS_NAMESPACE} \
pod-security.kubernetes.io/enforce=restricted \
pod-security.kubernetes.io/enforce-version=latest \
pod-security.kubernetes.io/warn=restricted \
pod-security.kubernetes.io/warn-version=latest \
--overwrite
# Delete NATS namespace
delete-namespace:
kubectl delete namespace ${NATS_NAMESPACE} --ignore-not-found
# Install NATS
install:
#!/bin/bash
set -euo pipefail
just create-namespace
just add-helm-repo
gomplate -f values.gomplate.yaml -o values.yaml
helm upgrade --install nats nats/nats \
--version ${NATS_CHART_VERSION} \
-n ${NATS_NAMESPACE} \
-f values.yaml \
--wait
echo ""
echo "NATS installed successfully"
echo "Namespace: ${NATS_NAMESPACE}"
echo "Replicas: ${NATS_REPLICAS}"
echo "JetStream enabled: ${NATS_JETSTREAM_ENABLED}"
echo ""
echo "Internal URL: nats://nats.${NATS_NAMESPACE}.svc:4222"
# Upgrade NATS
upgrade:
#!/bin/bash
set -euo pipefail
gomplate -f values.gomplate.yaml -o values.yaml
helm upgrade nats nats/nats \
--version ${NATS_CHART_VERSION} \
-n ${NATS_NAMESPACE} \
-f values.yaml \
--wait
echo "NATS upgraded successfully"
# Uninstall NATS
uninstall:
#!/bin/bash
set -euo pipefail
if ! gum confirm "Are you sure you want to uninstall NATS?"; then
echo "Aborted"
exit 0
fi
helm uninstall nats -n ${NATS_NAMESPACE} --wait --ignore-not-found
just delete-namespace
echo "NATS uninstalled"
# Show NATS status
status:
kubectl get pods -n ${NATS_NAMESPACE}
kubectl get svc -n ${NATS_NAMESPACE}
# Show NATS logs
logs:
kubectl logs -n ${NATS_NAMESPACE} -l app.kubernetes.io/name=nats -f
# Show server info via monitoring endpoint
server-info:
kubectl exec -n ${NATS_NAMESPACE} nats-0 -c nats -- \
wget -qO- http://localhost:8222/varz | head -50
# Show JetStream info via monitoring endpoint
js-info:
kubectl exec -n ${NATS_NAMESPACE} nats-0 -c nats -- \
wget -qO- http://localhost:8222/jsz
# Port forward for local testing
port-forward:
@echo "NATS available at localhost:4222"
@echo "Monitor available at http://localhost:8222"
kubectl port-forward -n ${NATS_NAMESPACE} svc/nats 4222:4222 8222:8222

64
nats/values.gomplate.yaml Normal file
View File

@@ -0,0 +1,64 @@
config:
cluster:
enabled: {{ if gt (conv.ToInt .Env.NATS_REPLICAS) 1 }}true{{ else }}false{{ end }}
replicas: {{ .Env.NATS_REPLICAS }}
jetstream:
enabled: {{ .Env.NATS_JETSTREAM_ENABLED }}
fileStore:
enabled: true
dir: /data
pvc:
enabled: true
size: {{ .Env.NATS_JETSTREAM_STORAGE_SIZE }}
memoryStore:
enabled: true
maxSize: {{ .Env.NATS_JETSTREAM_MEMORY_SIZE }}
monitor:
enabled: true
port: 8222
container:
resources:
requests:
cpu: 50m
memory: 64Mi
limits:
cpu: 500m
memory: 512Mi
merge:
securityContext:
runAsUser: 1000
runAsNonRoot: true
allowPrivilegeEscalation: false
seccompProfile:
type: RuntimeDefault
capabilities:
drop:
- ALL
reloader:
enabled: true
merge:
securityContext:
runAsUser: 1000
runAsNonRoot: true
allowPrivilegeEscalation: false
seccompProfile:
type: RuntimeDefault
capabilities:
drop:
- ALL
podTemplate:
merge:
spec:
securityContext:
fsGroup: 1000
seccompProfile:
type: RuntimeDefault
natsBox:
enabled: false

64
nats/values.yaml Normal file
View File

@@ -0,0 +1,64 @@
config:
cluster:
enabled: false
replicas: 1
jetstream:
enabled: true
fileStore:
enabled: true
dir: /data
pvc:
enabled: true
size: 10Gi
memoryStore:
enabled: true
maxSize: 256Mi
monitor:
enabled: true
port: 8222
container:
resources:
requests:
cpu: 50m
memory: 64Mi
limits:
cpu: 500m
memory: 512Mi
merge:
securityContext:
runAsUser: 1000
runAsNonRoot: true
allowPrivilegeEscalation: false
seccompProfile:
type: RuntimeDefault
capabilities:
drop:
- ALL
reloader:
enabled: true
merge:
securityContext:
runAsUser: 1000
runAsNonRoot: true
allowPrivilegeEscalation: false
seccompProfile:
type: RuntimeDefault
capabilities:
drop:
- ALL
podTemplate:
merge:
spec:
securityContext:
fsGroup: 1000
seccompProfile:
type: RuntimeDefault
natsBox:
enabled: false

View File

@@ -8,6 +8,11 @@ export OLLAMA_GPU_TYPE := env("OLLAMA_GPU_TYPE", "nvidia")
export OLLAMA_GPU_COUNT := env("OLLAMA_GPU_COUNT", "1")
export OLLAMA_MODELS := env("OLLAMA_MODELS", "")
export OLLAMA_STORAGE_SIZE := env("OLLAMA_STORAGE_SIZE", "30Gi")
export OLLAMA_HELM_TIMEOUT := env("OLLAMA_HELM_TIMEOUT", "60m")
export OLLAMA_MEMORY_REQUEST := env("OLLAMA_MEMORY_REQUEST", "2Gi")
export OLLAMA_MEMORY_LIMIT := env("OLLAMA_MEMORY_LIMIT", "12Gi")
export OLLAMA_CPU_REQUEST := env("OLLAMA_CPU_REQUEST", "25m")
export OLLAMA_CPU_LIMIT := env("OLLAMA_CPU_LIMIT", "100m")
[private]
default:
@@ -66,7 +71,8 @@ install:
gomplate -f values.gomplate.yaml -o values.yaml
helm upgrade --install ollama ollama/ollama \
--version ${OLLAMA_CHART_VERSION} -n ${OLLAMA_NAMESPACE} --wait \
--version ${OLLAMA_CHART_VERSION} -n ${OLLAMA_NAMESPACE} \
--wait --timeout ${OLLAMA_HELM_TIMEOUT} \
-f values.yaml
echo ""
@@ -97,7 +103,8 @@ upgrade:
gomplate -f values.gomplate.yaml -o values.yaml
helm upgrade ollama ollama/ollama \
--version ${OLLAMA_CHART_VERSION} -n ${OLLAMA_NAMESPACE} --wait \
--version ${OLLAMA_CHART_VERSION} -n ${OLLAMA_NAMESPACE} \
--wait --timeout ${OLLAMA_HELM_TIMEOUT} \
-f values.yaml
echo "Ollama upgraded successfully"

View File

@@ -36,11 +36,11 @@ securityContext:
resources:
requests:
cpu: 25m
memory: 2Gi
cpu: {{ .Env.OLLAMA_CPU_REQUEST }}
memory: {{ .Env.OLLAMA_MEMORY_REQUEST }}
limits:
cpu: 100m
memory: 8Gi
cpu: {{ .Env.OLLAMA_CPU_LIMIT }}
memory: {{ .Env.OLLAMA_MEMORY_LIMIT }}
persistentVolume:
enabled: true

1
temporal/.gitignore vendored Normal file
View File

@@ -0,0 +1 @@
temporal-values.yaml

401
temporal/README.md Normal file
View File

@@ -0,0 +1,401 @@
# Temporal
Durable workflow execution platform for building reliable distributed applications:
- **Durable Execution**: Workflows survive process and infrastructure failures
- **Language Support**: SDKs for Go, Java, Python, TypeScript, .NET, PHP
- **Visibility**: Query and observe workflow state via Web UI and APIs
- **Scalability**: Horizontally scalable architecture
- **Multi-tenancy**: Namespace-based isolation for workflows
## Prerequisites
- Kubernetes cluster (k3s)
- PostgreSQL cluster (CloudNativePG)
- Keycloak installed and configured
- Vault for secrets management
- External Secrets Operator (optional, for Vault integration)
## Installation
```bash
just temporal::install
```
You will be prompted for:
- **Temporal host (FQDN)**: e.g., `temporal.example.com`
- **Keycloak host (FQDN)**: e.g., `auth.example.com`
- **Enable Prometheus monitoring**: If kube-prometheus-stack is installed
### What Gets Installed
- Temporal Server (frontend, history, matching, worker services)
- Temporal Web UI with Keycloak OIDC authentication
- Temporal Admin Tools for cluster management
- PostgreSQL databases (`temporal`, `temporal_visibility`)
- Keycloak OAuth client (confidential client)
- Vault secrets (if External Secrets Operator is available)
## Configuration
Environment variables (set in `.env.local` or override):
| Variable | Default | Description |
| -------- | ------- | ----------- |
| `TEMPORAL_NAMESPACE` | `temporal` | Kubernetes namespace |
| `TEMPORAL_CHART_VERSION` | `0.52.0` | Helm chart version |
| `TEMPORAL_HOST` | (prompt) | External hostname (FQDN) |
| `TEMPORAL_OIDC_CLIENT_ID` | `temporal` | Keycloak client ID |
| `KEYCLOAK_HOST` | (prompt) | Keycloak hostname (FQDN) |
| `KEYCLOAK_REALM` | `buunstack` | Keycloak realm |
| `MONITORING_ENABLED` | (prompt) | Enable Prometheus ServiceMonitor |
## Architecture
```plain
External Users
|
Cloudflare Tunnel (HTTPS)
|
Traefik Ingress (HTTPS)
|
Temporal Web UI (HTTP inside cluster)
|-- OAuth --> Keycloak (authentication)
|
Temporal Server
|-- Frontend Service (gRPC :7233)
| |-- Client connections
| |-- Workflow/Activity APIs
|
|-- History Service
| |-- Workflow state management
| |-- Event sourcing
|
|-- Matching Service
| |-- Task queue management
| |-- Worker polling
|
|-- Worker Service
| |-- System workflows
| |-- Archival
|
PostgreSQL (temporal, temporal_visibility)
```
**Key Components**:
- **Frontend**: Entry point for all client requests (gRPC API)
- **History**: Maintains workflow execution history and state
- **Matching**: Routes tasks to appropriate workers
- **Worker**: Executes internal system workflows
- **Web UI**: Browser-based workflow monitoring and management
- **Admin Tools**: CLI tools for cluster administration
## Usage
### Access Web UI
1. Navigate to `https://your-temporal-host/`
2. Authenticate via Keycloak SSO
3. Select a namespace to view workflows
### Temporal CLI Setup (Local Development)
The Temporal gRPC endpoint is only accessible within the cluster network. Use [Telepresence](https://www.telepresence.io/) to connect from your local machine.
#### Step 1: Connect to the Cluster
```bash
telepresence connect
```
#### Step 2: Configure Temporal CLI
Set environment variables (add to `.bashrc`, `.zshrc`, or use direnv):
```bash
export TEMPORAL_ADDRESS="temporal-frontend.temporal:7233"
export TEMPORAL_NAMESPACE="default"
```
Or create a named environment for multiple clusters:
```bash
# Configure named environment
temporal env set --env buun -k address -v temporal-frontend.temporal:7233
temporal env set --env buun -k namespace -v default
# Use with commands
temporal workflow list --env buun
```
#### Step 3: Verify Connection
```bash
# Check telepresence status
telepresence status
# Test Temporal connection
temporal operator namespace list
```
#### CLI Examples
```bash
# List workflows
temporal workflow list
# Describe a workflow
temporal workflow describe --workflow-id my-workflow-id
# Query workflow state
temporal workflow query --workflow-id my-workflow-id --type my-query
# Signal a workflow
temporal workflow signal --workflow-id my-workflow-id --name my-signal
# Terminate a workflow
temporal workflow terminate --workflow-id my-workflow-id --reason "manual termination"
```
### Create a Temporal Namespace
Before running workflows, create a namespace:
```bash
just temporal::create-temporal-namespace default
```
With custom retention period:
```bash
just temporal::create-temporal-namespace myapp 7d
```
### List Temporal Namespaces
```bash
just temporal::list-temporal-namespaces
```
### Cluster Health Check
```bash
just temporal::cluster-info
```
### Connect Workers
Workers connect to the Temporal Frontend service. From within the cluster:
```text
temporal-frontend.temporal:7233
```
Example Python worker:
```python
from temporalio.client import Client
from temporalio.worker import Worker
async def main():
client = await Client.connect("temporal-frontend.temporal:7233")
worker = Worker(
client,
task_queue="my-task-queue",
workflows=[MyWorkflow],
activities=[my_activity],
)
await worker.run()
```
Example Go worker:
```go
import (
"go.temporal.io/sdk/client"
"go.temporal.io/sdk/worker"
)
func main() {
c, _ := client.Dial(client.Options{
HostPort: "temporal-frontend.temporal:7233",
})
defer c.Close()
w := worker.New(c, "my-task-queue", worker.Options{})
w.RegisterWorkflow(MyWorkflow)
w.RegisterActivity(MyActivity)
w.Run(worker.InterruptCh())
}
```
## Authentication
### Web UI (OIDC)
- Users authenticate via Keycloak
- Standard OIDC flow with Authorization Code grant
- Configured via environment variables in the Web UI deployment
### gRPC API
- By default, no authentication is required for gRPC connections within the cluster
- For production, configure mTLS or JWT-based authorization
## Management
### Upgrade Temporal
```bash
just temporal::upgrade
```
### Uninstall
```bash
just temporal::uninstall
```
This removes:
- Helm release and all Kubernetes resources
- Namespace
- Keycloak client
**Note**: The following resources are NOT deleted:
- PostgreSQL databases (`temporal`, `temporal_visibility`)
- Vault secrets
### Full Cleanup
To remove everything including databases and Vault secrets:
```bash
just temporal::uninstall true
```
Or manually:
```bash
just temporal::delete-postgres-user-and-db
```
## Troubleshooting
### Check Pod Status
```bash
kubectl get pods -n temporal
```
Expected pods:
- `temporal-frontend-*` - Frontend service
- `temporal-history-*` - History service
- `temporal-matching-*` - Matching service
- `temporal-worker-*` - Worker service
- `temporal-web-*` - Web UI
- `temporal-admintools-*` - Admin tools
### View Logs
```bash
# Frontend logs
kubectl logs -n temporal deployment/temporal-frontend --tail=100
# History logs
kubectl logs -n temporal deployment/temporal-history --tail=100
# Web UI logs
kubectl logs -n temporal deployment/temporal-web --tail=100
```
### Database Connection Issues
Check PostgreSQL connectivity:
```bash
kubectl exec -n temporal deployment/temporal-admintools -- \
psql -h postgres-cluster-rw.postgres -U temporal -d temporal -c "SELECT 1"
```
### Schema Issues
If schema initialization fails, check the schema job:
```bash
kubectl logs -n temporal -l app.kubernetes.io/component=schema --all-containers
```
### Service Discovery Issues
Verify services are running:
```bash
kubectl get svc -n temporal
```
Test frontend connectivity from admin tools:
```bash
kubectl exec -n temporal deployment/temporal-admintools -- \
tctl cluster health
```
### Web UI Login Issues
Verify Keycloak client configuration:
```bash
just keycloak::get-client buunstack temporal
```
Check Web UI environment variables:
```bash
kubectl get deployment temporal-web -n temporal -o jsonpath='{.spec.template.spec.containers[0].env}' | jq
```
## Configuration Files
| File | Description |
| ---- | ----------- |
| `temporal-values.gomplate.yaml` | Helm values template |
| `postgres-external-secret.gomplate.yaml` | PostgreSQL credentials ExternalSecret |
| `keycloak-auth-external-secret.gomplate.yaml` | Keycloak OIDC credentials ExternalSecret |
## Security Considerations
- **Pod Security Standards**: Namespace configured with **baseline** enforcement
- **Server Security**: Temporal server components run with restricted-compliant security contexts
### Why Not Restricted?
The namespace cannot use `restricted` Pod Security Standards due to the Temporal Web UI image (`temporalio/ui`):
- The image writes configuration files to `./config/docker.yaml` at startup
- The container's filesystem is owned by root (UID 0)
- When running as non-root user (UID 1000), the container cannot write to these paths
- Error: `unable to create open ./config/docker.yaml: permission denied`
The Temporal server components (frontend, history, matching, worker) **do** meet `restricted` requirements and run with full security hardening. Only the Web UI component requires `baseline`.
### Server Security Context
Temporal server components (frontend, history, matching, worker) run with:
- `runAsNonRoot: true`
- `runAsUser: 1000`
- `allowPrivilegeEscalation: false`
- `seccompProfile.type: RuntimeDefault`
- `capabilities.drop: [ALL]`
## References
- [Temporal Documentation](https://docs.temporal.io/)
- [Temporal GitHub](https://github.com/temporalio/temporal)
- [Temporal Helm Charts](https://github.com/temporalio/helm-charts)

434
temporal/justfile Normal file
View File

@@ -0,0 +1,434 @@
set fallback := true
export TEMPORAL_NAMESPACE := env("TEMPORAL_NAMESPACE", "temporal")
export TEMPORAL_CHART_VERSION := env("TEMPORAL_CHART_VERSION", "0.72.0")
export TEMPORAL_HOST := env("TEMPORAL_HOST", "")
export TEMPORAL_OIDC_CLIENT_ID := env("TEMPORAL_OIDC_CLIENT_ID", "temporal")
export EXTERNAL_SECRETS_NAMESPACE := env("EXTERNAL_SECRETS_NAMESPACE", "external-secrets")
export PROMETHEUS_NAMESPACE := env("PROMETHEUS_NAMESPACE", "monitoring")
export MONITORING_ENABLED := env("MONITORING_ENABLED", "")
export KEYCLOAK_REALM := env("KEYCLOAK_REALM", "buunstack")
export KEYCLOAK_HOST := env("KEYCLOAK_HOST", "")
export K8S_VAULT_NAMESPACE := env("K8S_VAULT_NAMESPACE", "vault")
[private]
default:
@just --list --unsorted --list-submodules
# Add Helm repository
add-helm-repo:
helm repo add temporal https://go.temporal.io/helm-charts
helm repo update temporal
# Remove Helm repository
remove-helm-repo:
helm repo remove temporal
# Create Temporal namespace
create-namespace:
kubectl get namespace ${TEMPORAL_NAMESPACE} &>/dev/null || \
kubectl create namespace ${TEMPORAL_NAMESPACE}
# Delete Temporal namespace
delete-namespace:
kubectl delete namespace ${TEMPORAL_NAMESPACE} --ignore-not-found
# Create PostgreSQL user and databases for Temporal
create-postgres-user-and-db:
#!/bin/bash
set -euo pipefail
if just postgres::user-exists temporal &>/dev/null; then
echo "PostgreSQL user 'temporal' already exists"
else
echo "Creating PostgreSQL user and databases..."
PG_PASSWORD=$(just utils::random-password)
just postgres::create-user-and-db temporal temporal "${PG_PASSWORD}"
just postgres::create-db temporal_visibility
just postgres::grant temporal_visibility temporal
just vault::put temporal/db username=temporal password="${PG_PASSWORD}"
echo "PostgreSQL user and databases created."
fi
# Delete PostgreSQL user and databases
delete-postgres-user-and-db:
#!/bin/bash
set -euo pipefail
if gum confirm "Delete PostgreSQL user and databases for Temporal?"; then
just postgres::delete-db temporal || true
just postgres::delete-db temporal_visibility || true
just postgres::delete-user temporal || true
just vault::delete temporal/db || true
echo "PostgreSQL user and databases deleted."
else
echo "Cancelled."
fi
# Create Postgres secret
create-postgres-secret:
#!/bin/bash
set -euo pipefail
if kubectl get secret temporal-postgres-auth -n ${TEMPORAL_NAMESPACE} &>/dev/null; then
echo "Postgres auth secret already exists"
exit 0
fi
if helm status external-secrets -n ${EXTERNAL_SECRETS_NAMESPACE} &>/dev/null; then
echo "External Secrets Operator detected. Creating ExternalSecret..."
kubectl delete externalsecret temporal-postgres-auth -n ${TEMPORAL_NAMESPACE} --ignore-not-found
gomplate -f postgres-external-secret.gomplate.yaml | kubectl apply -f -
echo "Waiting for ExternalSecret to sync..."
kubectl wait --for=condition=Ready externalsecret/temporal-postgres-auth \
-n ${TEMPORAL_NAMESPACE} --timeout=60s
else
echo "Creating Kubernetes Secret directly..."
PG_USERNAME=$(just vault::get temporal/db username)
PG_PASSWORD=$(just vault::get temporal/db password)
kubectl create secret generic temporal-postgres-auth \
--from-literal=password="${PG_PASSWORD}" \
-n ${TEMPORAL_NAMESPACE}
fi
# Delete Postgres secret
delete-postgres-secret:
kubectl delete externalsecret temporal-postgres-auth -n ${TEMPORAL_NAMESPACE} --ignore-not-found
kubectl delete secret temporal-postgres-auth -n ${TEMPORAL_NAMESPACE} --ignore-not-found
# Create Keycloak client for Temporal Web UI
create-keycloak-client:
#!/bin/bash
set -euo pipefail
while [ -z "${TEMPORAL_HOST}" ]; do
TEMPORAL_HOST=$(
gum input --prompt="Temporal host (FQDN): " --width=100 \
--placeholder="e.g., temporal.example.com"
)
done
echo "Creating Keycloak client for Temporal..."
just keycloak::delete-client ${KEYCLOAK_REALM} ${TEMPORAL_OIDC_CLIENT_ID} || true
CLIENT_SECRET=$(just utils::random-password)
just keycloak::create-client \
realm=${KEYCLOAK_REALM} \
client_id=${TEMPORAL_OIDC_CLIENT_ID} \
redirect_url="https://${TEMPORAL_HOST}/*" \
client_secret="${CLIENT_SECRET}"
kubectl delete secret temporal-oauth-temp -n ${TEMPORAL_NAMESPACE} --ignore-not-found
kubectl create secret generic temporal-oauth-temp -n ${TEMPORAL_NAMESPACE} \
--from-literal=client_id="${TEMPORAL_OIDC_CLIENT_ID}" \
--from-literal=client_secret="${CLIENT_SECRET}"
echo "Keycloak client created successfully"
echo "Client ID: ${TEMPORAL_OIDC_CLIENT_ID}"
echo "Redirect URI: https://${TEMPORAL_HOST}/*"
# Delete Keycloak client
delete-keycloak-client:
#!/bin/bash
set -euo pipefail
echo "Deleting Keycloak client for Temporal..."
just keycloak::delete-client ${KEYCLOAK_REALM} ${TEMPORAL_OIDC_CLIENT_ID} || true
kubectl delete secret temporal-oauth-temp -n ${TEMPORAL_NAMESPACE} --ignore-not-found
if just vault::exist keycloak/client/temporal &>/dev/null; then
just vault::delete keycloak/client/temporal
fi
# Create Keycloak auth secret
create-keycloak-auth-secret:
#!/bin/bash
set -euo pipefail
if kubectl get secret temporal-oauth-temp -n ${TEMPORAL_NAMESPACE} &>/dev/null; then
oauth_client_id=$(kubectl get secret temporal-oauth-temp -n ${TEMPORAL_NAMESPACE} \
-o jsonpath='{.data.client_id}' | base64 -d)
oauth_client_secret=$(kubectl get secret temporal-oauth-temp -n ${TEMPORAL_NAMESPACE} \
-o jsonpath='{.data.client_secret}' | base64 -d)
elif helm status vault -n ${K8S_VAULT_NAMESPACE} &>/dev/null && \
just vault::get keycloak/client/temporal client_secret &>/dev/null; then
oauth_client_id=$(just vault::get keycloak/client/temporal client_id)
oauth_client_secret=$(just vault::get keycloak/client/temporal client_secret)
else
echo "Error: Cannot retrieve OAuth client secret. Please run 'just temporal::create-keycloak-client' first."
exit 1
fi
if helm status external-secrets -n ${EXTERNAL_SECRETS_NAMESPACE} &>/dev/null; then
echo "External Secrets Operator detected. Storing secrets in Vault..."
just vault::put keycloak/client/temporal \
client_id="${oauth_client_id}" \
client_secret="${oauth_client_secret}"
kubectl delete secret temporal-web-auth -n ${TEMPORAL_NAMESPACE} --ignore-not-found
kubectl delete externalsecret temporal-web-auth -n ${TEMPORAL_NAMESPACE} --ignore-not-found
gomplate -f keycloak-auth-external-secret.gomplate.yaml | kubectl apply -f -
echo "Waiting for ExternalSecret to sync..."
kubectl wait --for=condition=Ready externalsecret/temporal-web-auth \
-n ${TEMPORAL_NAMESPACE} --timeout=60s
echo "ExternalSecret created successfully"
else
echo "External Secrets Operator not found. Creating Kubernetes Secret directly..."
kubectl delete secret temporal-web-auth -n ${TEMPORAL_NAMESPACE} --ignore-not-found
kubectl create secret generic temporal-web-auth -n ${TEMPORAL_NAMESPACE} \
--from-literal=TEMPORAL_AUTH_CLIENT_ID="${oauth_client_id}" \
--from-literal=TEMPORAL_AUTH_CLIENT_SECRET="${oauth_client_secret}"
if helm status vault -n ${K8S_VAULT_NAMESPACE} &>/dev/null; then
just vault::put keycloak/client/temporal \
client_id="${oauth_client_id}" \
client_secret="${oauth_client_secret}"
fi
echo "Kubernetes Secret created successfully"
fi
kubectl delete secret temporal-oauth-temp -n ${TEMPORAL_NAMESPACE} --ignore-not-found
# Delete Keycloak auth secret
delete-keycloak-auth-secret:
kubectl delete externalsecret temporal-web-auth -n ${TEMPORAL_NAMESPACE} --ignore-not-found
kubectl delete secret temporal-web-auth -n ${TEMPORAL_NAMESPACE} --ignore-not-found
# Initialize Temporal database schema
init-schema:
#!/bin/bash
set -euo pipefail
echo "Initializing Temporal database schema..."
PG_HOST="postgres-cluster-rw.postgres"
PG_PORT="5432"
PG_USER=$(just vault::get temporal/db username)
PG_PASSWORD=$(just vault::get temporal/db password)
POD_NAME=$(kubectl get pods -n ${TEMPORAL_NAMESPACE} -l app.kubernetes.io/name=temporal-admintools \
-o jsonpath='{.items[0].metadata.name}' 2>/dev/null || echo "")
if [ -z "${POD_NAME}" ]; then
echo "Admin tools pod not found. Running schema setup job..."
exit 0
fi
echo "Setting up main database schema..."
kubectl exec -n ${TEMPORAL_NAMESPACE} ${POD_NAME} -- \
temporal-sql-tool --plugin postgres12 \
--endpoint ${PG_HOST} --port ${PG_PORT} \
--user ${PG_USER} --password ${PG_PASSWORD} \
--database temporal \
setup-schema -v 0.0
kubectl exec -n ${TEMPORAL_NAMESPACE} ${POD_NAME} -- \
temporal-sql-tool --plugin postgres12 \
--endpoint ${PG_HOST} --port ${PG_PORT} \
--user ${PG_USER} --password ${PG_PASSWORD} \
--database temporal \
update-schema -d /etc/temporal/schema/postgresql/v12/temporal/versioned
echo "Setting up visibility database schema..."
kubectl exec -n ${TEMPORAL_NAMESPACE} ${POD_NAME} -- \
temporal-sql-tool --plugin postgres12 \
--endpoint ${PG_HOST} --port ${PG_PORT} \
--user ${PG_USER} --password ${PG_PASSWORD} \
--database temporal_visibility \
setup-schema -v 0.0
kubectl exec -n ${TEMPORAL_NAMESPACE} ${POD_NAME} -- \
temporal-sql-tool --plugin postgres12 \
--endpoint ${PG_HOST} --port ${PG_PORT} \
--user ${PG_USER} --password ${PG_PASSWORD} \
--database temporal_visibility \
update-schema -d /etc/temporal/schema/postgresql/v12/visibility/versioned
echo "Schema initialization complete."
# Install Temporal
install:
#!/bin/bash
set -euo pipefail
while [ -z "${TEMPORAL_HOST}" ]; do
TEMPORAL_HOST=$(gum input --prompt="Temporal host (FQDN): " --width=80 \
--placeholder="e.g., temporal.example.com")
done
while [ -z "${KEYCLOAK_HOST}" ]; do
KEYCLOAK_HOST=$(gum input --prompt="Keycloak host (FQDN): " --width=80 \
--placeholder="e.g., auth.example.com")
done
if helm status kube-prometheus-stack -n ${PROMETHEUS_NAMESPACE} &>/dev/null; then
if [ -z "${MONITORING_ENABLED}" ]; then
if gum confirm "Enable Prometheus monitoring?"; then
MONITORING_ENABLED="true"
fi
fi
fi
echo "Installing Temporal..."
just create-namespace
kubectl label namespace ${TEMPORAL_NAMESPACE} \
pod-security.kubernetes.io/enforce=baseline --overwrite
if [ "${MONITORING_ENABLED}" = "true" ]; then
kubectl label namespace ${TEMPORAL_NAMESPACE} \
buun.channel/enable-monitoring=true --overwrite
fi
echo "Setting up PostgreSQL database..."
just create-postgres-user-and-db
just create-postgres-secret
echo "Setting up Keycloak OIDC authentication..."
just create-keycloak-client
just create-keycloak-auth-secret
echo "Generating Helm values..."
just add-helm-repo
gomplate -f temporal-values.gomplate.yaml -o temporal-values.yaml
echo "Installing Temporal Helm chart..."
helm upgrade --cleanup-on-fail --install temporal temporal/temporal \
--version ${TEMPORAL_CHART_VERSION} -n ${TEMPORAL_NAMESPACE} --wait \
-f temporal-values.yaml --timeout 15m
echo "Configuring dynamic config for Worker Insights..."
kubectl patch configmap temporal-dynamic-config -n ${TEMPORAL_NAMESPACE} --type merge -p '{
"data": {
"dynamic_config.yaml": "frontend.WorkerHeartbeatsEnabled:\n - value: true\nfrontend.ListWorkersEnabled:\n - value: true\n"
}
}'
echo "Restarting frontend to apply dynamic config..."
kubectl rollout restart deployment temporal-frontend -n ${TEMPORAL_NAMESPACE}
kubectl rollout status deployment temporal-frontend -n ${TEMPORAL_NAMESPACE} --timeout=120s
echo ""
echo "Temporal installed successfully!"
echo "Access Temporal Web UI at: https://${TEMPORAL_HOST}"
echo ""
echo "OIDC authentication is configured with Keycloak."
echo "Users can login with their Keycloak credentials."
# Upgrade Temporal
upgrade:
#!/bin/bash
set -euo pipefail
while [ -z "${TEMPORAL_HOST}" ]; do
TEMPORAL_HOST=$(gum input --prompt="Temporal host (FQDN): " --width=80)
done
while [ -z "${KEYCLOAK_HOST}" ]; do
KEYCLOAK_HOST=$(gum input --prompt="Keycloak host (FQDN): " --width=80)
done
if helm status kube-prometheus-stack -n ${PROMETHEUS_NAMESPACE} &>/dev/null; then
if [ -z "${MONITORING_ENABLED}" ]; then
if gum confirm "Enable Prometheus monitoring?"; then
MONITORING_ENABLED="true"
fi
fi
fi
if [ "${MONITORING_ENABLED}" = "true" ]; then
kubectl label namespace ${TEMPORAL_NAMESPACE} \
buun.channel/enable-monitoring=true --overwrite
fi
echo "Upgrading Temporal..."
gomplate -f temporal-values.gomplate.yaml -o temporal-values.yaml
helm upgrade temporal temporal/temporal \
--version ${TEMPORAL_CHART_VERSION} -n ${TEMPORAL_NAMESPACE} --wait \
-f temporal-values.yaml --timeout 15m
echo "Configuring dynamic config for Worker Insights..."
kubectl patch configmap temporal-dynamic-config -n ${TEMPORAL_NAMESPACE} --type merge -p '{
"data": {
"dynamic_config.yaml": "frontend.WorkerHeartbeatsEnabled:\n - value: true\nfrontend.ListWorkersEnabled:\n - value: true\n"
}
}'
echo "Restarting frontend to apply dynamic config..."
kubectl rollout restart deployment temporal-frontend -n ${TEMPORAL_NAMESPACE}
kubectl rollout status deployment temporal-frontend -n ${TEMPORAL_NAMESPACE} --timeout=120s
echo ""
echo "Temporal upgraded successfully!"
echo "Access Temporal Web UI at: https://${TEMPORAL_HOST}"
# Uninstall Temporal (delete-data: true to delete database and Vault secrets)
uninstall delete-data='false':
#!/bin/bash
set -euo pipefail
if ! gum confirm "Uninstall Temporal?"; then
echo "Cancelled."
exit 0
fi
echo "Uninstalling Temporal..."
helm uninstall temporal -n ${TEMPORAL_NAMESPACE} --ignore-not-found --wait
just delete-keycloak-auth-secret || true
just delete-keycloak-client || true
just delete-postgres-secret
just delete-namespace
if [ "{{ delete-data }}" = "true" ]; then
echo "Deleting database and Vault secrets..."
just postgres::delete-db temporal || true
just postgres::delete-db temporal_visibility || true
just postgres::delete-user temporal || true
just vault::delete temporal/db || true
just vault::delete keycloak/client/temporal || true
echo "Temporal uninstalled with all data deleted."
else
echo "Temporal uninstalled."
echo ""
echo "Note: The following resources were NOT deleted:"
echo " - PostgreSQL user and databases (temporal, temporal_visibility)"
echo " - Vault secrets (temporal/db, keycloak/client/temporal)"
echo ""
echo "To delete all data, run:"
echo " just temporal::uninstall true"
fi
# Create a Temporal namespace (workflow namespace, not Kubernetes)
create-temporal-namespace name='' retention='3d':
#!/bin/bash
set -euo pipefail
name="{{ name }}"
retention="{{ retention }}"
while [ -z "${name}" ]; do
name=$(gum input --prompt="Namespace name: " --width=80 --placeholder="e.g., default")
done
POD_NAME=$(kubectl get pods -n ${TEMPORAL_NAMESPACE} -l app.kubernetes.io/name=temporal-admintools \
-o jsonpath='{.items[0].metadata.name}')
kubectl exec -n ${TEMPORAL_NAMESPACE} ${POD_NAME} -- \
tctl --namespace "${name}" namespace register --retention "${retention}"
echo "Namespace '${name}' created with retention ${retention}."
# List Temporal namespaces
list-temporal-namespaces:
#!/bin/bash
set -euo pipefail
POD_NAME=$(kubectl get pods -n ${TEMPORAL_NAMESPACE} -l app.kubernetes.io/name=temporal-admintools \
-o jsonpath='{.items[0].metadata.name}')
kubectl exec -n ${TEMPORAL_NAMESPACE} ${POD_NAME} -- tctl namespace list
# Get Temporal cluster info
cluster-info:
#!/bin/bash
set -euo pipefail
POD_NAME=$(kubectl get pods -n ${TEMPORAL_NAMESPACE} -l app.kubernetes.io/name=temporal-admintools \
-o jsonpath='{.items[0].metadata.name}')
kubectl exec -n ${TEMPORAL_NAMESPACE} ${POD_NAME} -- tctl cluster health

View File

@@ -0,0 +1,22 @@
apiVersion: external-secrets.io/v1
kind: ExternalSecret
metadata:
name: temporal-web-auth
namespace: {{ .Env.TEMPORAL_NAMESPACE }}
spec:
refreshInterval: 1h
secretStoreRef:
name: vault-secret-store
kind: ClusterSecretStore
target:
name: temporal-web-auth
creationPolicy: Owner
data:
- secretKey: TEMPORAL_AUTH_CLIENT_ID
remoteRef:
key: keycloak/client/temporal
property: client_id
- secretKey: TEMPORAL_AUTH_CLIENT_SECRET
remoteRef:
key: keycloak/client/temporal
property: client_secret

View File

@@ -0,0 +1,18 @@
apiVersion: external-secrets.io/v1
kind: ExternalSecret
metadata:
name: temporal-postgres-auth
namespace: {{ .Env.TEMPORAL_NAMESPACE }}
spec:
refreshInterval: 1h
secretStoreRef:
name: vault-secret-store
kind: ClusterSecretStore
target:
name: temporal-postgres-auth
creationPolicy: Owner
data:
- secretKey: password
remoteRef:
key: temporal/db
property: password

View File

@@ -0,0 +1,204 @@
server:
replicaCount: 1
securityContext:
runAsUser: 1000
runAsGroup: 1000
fsGroup: 1000
runAsNonRoot: true
seccompProfile:
type: RuntimeDefault
config:
persistence:
default:
driver: "sql"
sql:
driver: "postgres12"
host: "postgres-cluster-rw.postgres"
port: 5432
database: temporal
user: temporal
existingSecret: temporal-postgres-auth
maxConns: 20
maxIdleConns: 20
maxConnLifetime: "1h"
visibility:
driver: "sql"
sql:
driver: "postgres12"
host: "postgres-cluster-rw.postgres"
port: 5432
database: temporal_visibility
user: temporal
existingSecret: temporal-postgres-auth
maxConns: 20
maxIdleConns: 20
maxConnLifetime: "1h"
{{- if .Env.MONITORING_ENABLED }}
metrics:
serviceMonitor:
enabled: true
additionalLabels:
release: kube-prometheus-stack
{{- end }}
frontend:
containerSecurityContext:
allowPrivilegeEscalation: false
capabilities:
drop:
- ALL
readOnlyRootFilesystem: false
resources:
requests:
cpu: 100m
memory: 128Mi
limits:
cpu: 500m
memory: 512Mi
history:
containerSecurityContext:
allowPrivilegeEscalation: false
capabilities:
drop:
- ALL
readOnlyRootFilesystem: false
resources:
requests:
cpu: 100m
memory: 128Mi
limits:
cpu: 500m
memory: 512Mi
matching:
containerSecurityContext:
allowPrivilegeEscalation: false
capabilities:
drop:
- ALL
readOnlyRootFilesystem: false
resources:
requests:
cpu: 100m
memory: 128Mi
limits:
cpu: 500m
memory: 512Mi
worker:
containerSecurityContext:
allowPrivilegeEscalation: false
capabilities:
drop:
- ALL
readOnlyRootFilesystem: false
resources:
requests:
cpu: 100m
memory: 128Mi
limits:
cpu: 500m
memory: 512Mi
admintools:
enabled: true
securityContext:
runAsUser: 1000
runAsGroup: 1000
fsGroup: 1000
runAsNonRoot: true
seccompProfile:
type: RuntimeDefault
containerSecurityContext:
allowPrivilegeEscalation: false
capabilities:
drop:
- ALL
readOnlyRootFilesystem: false
resources:
requests:
cpu: 50m
memory: 64Mi
limits:
cpu: 200m
memory: 256Mi
web:
enabled: true
replicaCount: 1
service:
type: ClusterIP
port: 8080
ingress:
enabled: true
className: traefik
annotations:
traefik.ingress.kubernetes.io/router.entrypoints: websecure
traefik.ingress.kubernetes.io/router.tls: "true"
hosts:
- {{ .Env.TEMPORAL_HOST }}
tls:
- secretName: temporal-web-tls
hosts:
- {{ .Env.TEMPORAL_HOST }}
additionalEnv:
- name: TEMPORAL_AUTH_ENABLED
value: "true"
- name: TEMPORAL_AUTH_PROVIDER_URL
value: "https://{{ .Env.KEYCLOAK_HOST }}/realms/{{ .Env.KEYCLOAK_REALM }}"
- name: TEMPORAL_AUTH_SCOPES
value: "openid,profile,email"
- name: TEMPORAL_AUTH_CALLBACK_URL
value: "https://{{ .Env.TEMPORAL_HOST }}/auth/sso/callback"
additionalEnvSecretName: temporal-web-auth
resources:
requests:
cpu: 50m
memory: 64Mi
limits:
cpu: 200m
memory: 256Mi
cassandra:
enabled: false
mysql:
enabled: false
postgresql:
enabled: false
elasticsearch:
enabled: false
prometheus:
enabled: false
grafana:
enabled: false
schema:
createDatabase:
enabled: false
setup:
enabled: true
backoffLimit: 100
update:
enabled: true
backoffLimit: 100
securityContext:
runAsUser: 1000
runAsGroup: 1000
fsGroup: 1000
runAsNonRoot: true
seccompProfile:
type: RuntimeDefault
containerSecurityContext:
allowPrivilegeEscalation: false
capabilities:
drop:
- ALL
readOnlyRootFilesystem: false