Skip to main content

​
Deployment Guide

Deploy Simili Bot to production with high reliability and performance.

​
Deployment Options

GitHub Action

Event-driven - runs on issue events

Docker

Container - self-hosted

Kubernetes

Orchestrated - enterprise scale
Simplest deployment - runs automatically on issue events. Pros:
  • Zero infrastructure
  • Automatic scaling
  • GitHub-native
  • Easy to debug
Cons:
  • Limited to GitHub
  • Per-run time limits
See GitHub Action Setup for details.

​
Self-Hosted Docker

Full control over deployment. Pros:
  • Control environment
  • Batch processing
  • Custom workflows
  • Integration with other tools
Cons:
  • Infrastructure management
  • Scaling complexity
  • Security responsibility

​
Kubernetes Deployment (Enterprise)

For large-scale deployments: Example: 
apiVersion: apps/v1
kind: Deployment
metadata:
  name: simili-bot
spec:
  replicas: 3
  selector:
    matchLabels:
      app: simili-bot
  template:
    metadata:
      labels:
        app: simili-bot
    spec:
      containers:
      - name: simili-bot
        image: similigh/simili-bot:v0.1.0
        env:
        - name: QDRANT_URL
          valueFrom:
            secretKeyRef:
              name: simili-secrets
              key: qdrant-url
        - name: QDRANT_API_KEY
          valueFrom:
            secretKeyRef:
              name: simili-secrets
              key: qdrant-api-key
        - name: GEMINI_API_KEY
          valueFrom:
            secretKeyRef:
              name: simili-secrets
              key: gemini-api-key
        - name: GITHUB_TOKEN
          valueFrom:
            secretKeyRef:
              name: simili-secrets
              key: github-token
        resources:
          requests:
            memory: "256Mi"
            cpu: "100m"
          limits:
            memory: "512Mi"
            cpu: "500m"
        volumeMounts:
        - name: config
          mountPath: /etc/simili
      volumes:
      - name: config
        configMap:
          name: simili-config

​
Production Checklist

  • Secrets configured securely
  • Error logging enabled
  • Rate limiting configured
  • Backup plan for vector DB
  • Monitoring and alerts set up
  • Load testing completed
  • Disaster recovery plan
  • Team trained on operations

​
Monitoring

Monitor these metrics:
  • API response times
  • Error rates
  • Vector DB size
  • GitHub API quota usage
  • Gemini API costs

​
Backup & Recovery

​
Qdrant Backup

Regularly backup your vector database: Qdrant Cloud:
  • Automatic backups included
  • Check backup schedule in dashboard
Self-Hosted: 
docker run --rm -v qdrant_storage:/qdrant/storage \
  -v backups:/backup \
  qdrant/qdrant \
  backup create /backup

​
Cost Optimization

  • Use dev cluster for testing
  • Monitor API usage
  • Archive old issues
  • Optimize batch sizes

​
Security

  • Use GitHub Actions secrets
  • Rotate API keys regularly
  • Limit GitHub token permissions
  • Use VPC for self-hosted
  • Enable TLS everywhere

​
Next Steps

Troubleshooting

Resolve common issues