test: CI/CD pipeline staging test - Redis aktiviert, Bad Gateway dokumentiert
This commit is contained in:
@@ -411,17 +411,54 @@ docker compose logs redis
|
||||
|
||||
### Performance Issues
|
||||
|
||||
```bash
|
||||
# Check MySQL slow queries
|
||||
docker exec gitea-mysql tail -100 /var/log/mysql/slow-queries.log
|
||||
**If Gitea has frequent outages or connection issues:**
|
||||
|
||||
# Analyze MySQL performance
|
||||
docker exec gitea-mysql mysql -u root -p$MYSQL_ROOT_PASSWORD \
|
||||
-e "SHOW PROCESSLIST;"
|
||||
1. **Update Gitea Configuration** (Recommended):
|
||||
```bash
|
||||
cd deployment/ansible
|
||||
ansible-playbook -i inventory/production.yml \
|
||||
playbooks/update-gitea-config.yml \
|
||||
--vault-password-file secrets/.vault_pass
|
||||
```
|
||||
|
||||
This playbook will:
|
||||
- Enable Redis cache for better performance and persistence
|
||||
- Configure database connection pooling
|
||||
- Set connection limits to prevent "Connection reset by peer" errors
|
||||
|
||||
# Check Redis memory usage
|
||||
docker exec gitea-redis redis-cli -a $REDIS_PASSWORD INFO memory
|
||||
```
|
||||
2. **Manual Troubleshooting**:
|
||||
```bash
|
||||
# Check PostgreSQL slow queries
|
||||
docker exec gitea-postgres psql -U gitea -d gitea -c "SELECT * FROM pg_stat_activity;"
|
||||
|
||||
# Check container resource usage
|
||||
docker stats gitea gitea-postgres gitea-redis
|
||||
|
||||
# Check Gitea logs for errors
|
||||
docker compose logs --tail 100 gitea | grep -i error
|
||||
|
||||
# Check Redis connection
|
||||
docker exec gitea-redis redis-cli -a $REDIS_PASSWORD ping
|
||||
```
|
||||
|
||||
### Known Issues
|
||||
|
||||
**Bad Gateway after many rapid requests (15-20 reloads):**
|
||||
- **Status**: Known issue, non-critical
|
||||
- **Symptoms**: Gitea returns "Bad Gateway" after 15-20 rapid page reloads, recovers after a few seconds
|
||||
- **Impact**: Low - Gitea is functional for normal usage
|
||||
- **Possible causes**:
|
||||
- Container restart during high load
|
||||
- Connection pool exhaustion (mitigated with increased limits)
|
||||
- Traefik service discovery delay in host network mode
|
||||
- **Workarounds**:
|
||||
- Wait a few seconds and retry
|
||||
- Use Redis cache (already enabled) for better performance
|
||||
- Consider adding rate limiting if needed (see Traefik middlewares)
|
||||
- **Future improvements**:
|
||||
- Monitor and optimize connection pool usage
|
||||
- Consider adding rate limiting middleware for Gitea
|
||||
- Investigate Traefik service discovery in host network mode
|
||||
|
||||
### Reset Admin Password
|
||||
|
||||
|
||||
Reference in New Issue
Block a user