Error Codes Reference
Error Codes Reference
Common error codes and solutions across the LLM platform.
Agent Errors
AGT-001: Agent Initialization Failed
Cause: Agent failed to initialize properly.
Solution:
# Check manifest buildkit ossa validate .agents/<agent>/manifest.json # Verify dependencies buildkit agents check-deps <agent-id> # Restart agent buildkit agents restart <agent-id>
AGT-002: Task Execution Timeout
Cause: Task exceeded maximum execution time.
Solution:
# Increase timeout in manifest # .agents/<agent>/manifest.json { "execution": { "timeout": 300000 // 5 minutes } }
AGT-003: OSSA Protocol Violation
Cause: Agent not compliant with OSSA specification.
Solution:
# Validate compliance buildkit ossa validate .agents/<agent>/manifest.json # Fix violations and redeploy buildkit agents deploy <agent-id>
LLM Errors
LLM-001: API Rate Limit Exceeded
Cause: Too many API requests.
Solution:
# Check rate limits in Phoenix open http://localhost:6006 # Implement backoff # Add to agent config { "llm": { "rateLimit": { "maxRequests": 50, "perMinute": 1 } } }
LLM-002: Invalid API Key
Cause: Missing or invalid API key.
Solution:
# Check token cat ~/.tokens/anthropic # Set token echo "your-api-key" > ~/.tokens/anthropic chmod 600 ~/.tokens/anthropic
LLM-003: Context Length Exceeded
Cause: Prompt too long for model.
Solution:
- Reduce prompt length
- Use Claude Opus (200k context)
- Implement chunking strategy
Database Errors
DB-001: Connection Failed
Cause: Cannot connect to database.
Solution:
# Check PostgreSQL status docker ps | grep postgres # Test connection psql -h localhost -U postgres -d llm_platform # Verify environment variables echo $DATABASE_URL
DB-002: Migration Failed
Cause: Database migration error.
Solution:
# Check migration status npm run migrate:status # Rollback and retry npm run migrate:rollback npm run migrate:up
Network Errors
NET-001: Service Unreachable
Cause: Cannot connect to service.
Solution: See Connection Issues
NET-002: DNS Resolution Failed
Cause: Cannot resolve hostname.
Solution: See Connection Issues
Kubernetes Errors
K8S-001: Pod OOMKilled
Cause: Pod exceeded memory limit.
Solution: See Kubernetes Troubleshooting
K8S-002: Image Pull Error
Cause: Cannot pull container image.
Solution:
# Check image exists docker pull <image> # Check registry credentials kubectl get secret regcred -n agents -o yaml # Create secret if missing kubectl create secret docker-registry regcred \ --docker-server=<registry> \ --docker-username=<user> \ --docker-password=<token> \ -n agents
Testing Errors
TEST-001: TDD Compliance Failed
Cause: Tests don't meet TDD requirements.
Solution: See TDD Enforcement
TEST-002: Coverage Below Threshold
Cause: Code coverage < 80%.
Solution: See Coverage Requirements
Getting Help
Check Logs
# Agent logs buildkit agents logs <agent-id> # Service logs docker logs <service> # Kubernetes logs kubectl logs -n <namespace> <pod>
View Traces
# Phoenix open http://localhost:6006 # Jaeger open http://localhost:16686
Report Issues
Create issue: https://gitlab.com/blueflyio/agent-platform/documentation/-/issues