# Find what's using the port (Linux/macOS)
lsof -i:8080
lsof -i:8082
lsof -i:8285
lsof -i:8091

# Find what's using the port (Windows)
netstat -ano | findstr:8080
taskkill /PID <PID> /F

# Or change ports in.env
HTTP_PORT=8081
RESPONSE_API_PORT=8083
MEDIA_API_PORT=8286

# Check the logs
make logs-api
make logs-mcp

# Or use Docker directly
docker logs <container-id>

# Check your .env file has all required values
cat .env

# Make sure database is running
make health-check

# Restart the service
docker restart <container-name>

# Verify all services are running
docker ps

# Check network connectivity
docker network ls
docker network inspect jan-server_default

# Test connectivity between services
docker exec llm-api curl http://media-api:8285/healthz

# Verify DNS resolution
docker exec llm-api nslookup media-api

# Check if PostgreSQL is running
docker ps | grep postgres

# Check database credentials in.env
cat.env | grep DATABASE

# Test connection
docker exec api-db psql -U jan_user -d jan_llm_api -c "SELECT 1"

# Verify database exists
docker exec api-db psql -U postgres -l | grep jan

# Create database
docker exec api-db psql -U postgres -c "CREATE DATABASE jan_llm_api"

# Or run migrations
docker exec llm-api /app/llm-api migrate

# Or with make
make db-migrate

# View migrations
docker exec api-db psql -U jan_user -d jan_llm_api \
 -c "SELECT name, version FROM schema_migrations"

# Reset database (WARNING: destroys data!)
make db-reset

# Re-migrate
make db-migrate

# Check disk usage
df -h

# Clean up Docker volumes
docker system prune -a --volumes

# Or manually remove volume
docker volume ls
docker volume rm <volume-name>

# Get a guest token
curl -X POST http://localhost:8000/llm/auth/guest-login

# Use token in requests
curl -H "Authorization: Bearer <token>" \
 http://localhost:8000/v1/models

# Check Keycloak is running
docker ps | grep keycloak

# Verify token is valid
jwt decode <token> # requires jwt-cli

# Verify service is running and healthy
curl http://localhost:8080/healthz # LLM API
curl http://localhost:8082/healthz # Response API
curl http://localhost:8285/healthz # Media API
curl http://localhost:8091/healthz # MCP Tools

# Verify Kong is routing correctly
curl http://localhost:8000/ # Kong health
curl http://localhost:8000/v1/models # Via Kong

# Increase timeout in .env
TOOL_EXECUTION_TIMEOUT=120s
MEDIA_REMOTE_FETCH_TIMEOUT=30s

# Check service performance
docker stats llm-api media-api response-api mcp-tools

# Look for stuck processes
make logs-api | grep -i timeout

# View detailed logs
docker logs <service-name> --tail=100 -f

# Check if service crashed
docker inspect <service-name>

# Restart service
docker restart <service-name>

# Or full restart
make down && make up-full

# Check if Keycloak is running
docker ps | grep keycloak

# Verify it's accessible
curl http://localhost:8085/admin

# Check logs
docker logs keycloak

# Restart Keycloak
docker restart keycloak

# Get new token
curl -X POST http://localhost:8000/llm/auth/guest-login

# Check token expiration
jwt decode <token>

# For user auth, check credentials
curl -X POST http://localhost:8085/auth/realms/jan/protocol/openid-connect/token \
 -d "client_id=llm-api&grant_type=password&username=admin&password=admin"

# Check memory usage
docker stats

# Increase Docker memory limit
# In Docker Desktop: Settings > Resources > Memory

# Or reduce services running
make down

# Clean up unused images and volumes
docker system prune -a --volumes

# Remove old containers
docker container prune

# Check image sizes
docker images --format "table {{.Repository}}\t{{.Size}}"

# Verify network exists
docker network ls

# Check network configuration
docker network inspect jan-server_default

# Recreate network if needed
docker network rm jan-server_default
docker network create jan-server_default

# Check events
kubectl describe pod -n jan-server <pod-name>

# Check node resources
kubectl top nodes

# Check available storage
kubectl get pvc -n jan-server

# Verify image exists
minikube image ls | grep jan

# Rebuild image
cd services/llm-api
docker build -t jan/llm-api:latest.
minikube image load jan/llm-api:latest

# Or update imagePullPolicy in values.yaml
imagePullPolicy: Never # For minikube

# Check service exists
kubectl get svc -n jan-server

# Port forward for access
kubectl port-forward -n jan-server svc/jan-server-llm-api 8080:8080

# Check service endpoints
kubectl get endpoints -n jan-server

# Monitor memory
docker stats llm-api

# Reduce batch size or concurrency
# Check service configuration in docker compose

# Look for memory leaks in logs
docker logs llm-api | grep -i memory

# Monitor CPU
docker stats

# Reduce concurrent requests
# Set rate limits in configuration

# Check for infinite loops or busy waits
docker logs llm-api | grep -i error

# Check database performance
docker exec api-db psql -U jan_user -d jan_llm_api \
 -c "SELECT query, calls, total_time FROM pg_stat_statements ORDER BY total_time DESC LIMIT 10"

# Enable query logging
# Set LOG_LEVEL=debug in.env

# Use monitoring stack for traces
make monitor-up
# Visit http://localhost:16686 (Jaeger)

# System info
docker version
docker-compose version
go version
kubectl version

# Service status
make health-check

# Logs from all services
make logs > debug-logs.txt

# Configuration (without secrets)
cat.env | grep -v _KEY | grep -v _PASSWORD > config.txt

# Docker system status
docker system df
docker ps -a