ljwx-k8s/gitea-actions-exporter

Gitea Actions Queue Exporter

Prometheus exporter for Gitea Actions queue metrics, designed for KEDA autoscaling of Gitea Action Runners.

Features

• Prometheus Metrics: Expose queue length, running jobs, and oldest job age
• Lightweight: Python-based with minimal dependencies
• Configurable: Environment variable configuration
• Health Checks: Built-in health check endpoint

Metrics

Metric	Type	Description
gitea_actions_pending_jobs	Gauge	Number of pending/waiting jobs
gitea_actions_running_jobs	Gauge	Number of running jobs
gitea_actions_total_jobs	Gauge	Total number of jobs
gitea_actions_queue_oldest_job_age_seconds	Gauge	Age of oldest pending job (seconds)
gitea_exporter_scrape_errors_total	Counter	Total scrape errors
gitea_exporter_last_scrape_timestamp	Gauge	Last successful scrape timestamp

Environment Variables

Variable	Default	Description
GITEA_URL	http://192.168.1.83:33000	Gitea server URL
GITEA_TOKEN	-	Gitea API token (optional for public repos)
EXPORTER_PORT	9999	Exporter listening port
SCRAPE_INTERVAL	15	Metrics scrape interval (seconds)

Build

docker build -t harbor.omniverseai.net:8080/library/gitea-actions-exporter:latest .
docker push harbor.omniverseai.net:8080/library/gitea-actions-exporter:latest

Run Locally

# With Python
python exporter.py

# With Docker
docker run -p 9999:9999 \
  -e GITEA_URL=http://192.168.1.83:33000 \
  gitea-actions-exporter:latest

Deploy to Kubernetes

kubectl apply -f k8s-manifests/

Usage with KEDA

triggers:
  - type: prometheus
    metadata:
      serverAddress: http://monitoring-kube-prometheus-prometheus.observability.svc.cluster.local:9090
      threshold: '5'
      query: sum(gitea_actions_pending_jobs{instance="http://192.168.1.83:33000"})

Development

# Install dependencies
pip install -r requirements.txt

# Run exporter
python exporter.py

# Access metrics
curl http://localhost:9999/metrics