mosaic/stack
1
0
Fork 0
Code Issues 10 Pull Requests Actions Packages Projects Releases 2 Wiki Activity
Files
e59e517d5cf28780052fe2016c7eb7349532b82e
stack/docs/scratchpads/orch-109-lifecycle.md
Jason Woltje 12abdfe81d feat(#93): implement agent spawn via federation 
Implements FED-010: Agent Spawn via Federation feature that enables
spawning and managing Claude agents on remote federated Mosaic Stack
instances via COMMAND message type.

Features:
- Federation agent command types (spawn, status, kill)
- FederationAgentService for handling agent operations
- Integration with orchestrator's agent spawner/lifecycle services
- API endpoints for spawning, querying status, and killing agents
- Full command routing through federation COMMAND infrastructure
- Comprehensive test coverage (12/12 tests passing)

Architecture:
- Hub → Spoke: Spawn agents on remote instances
- Command flow: FederationController → FederationAgentService →
  CommandService → Remote Orchestrator
- Response handling: Remote orchestrator returns agent status/results
- Security: Connection validation, signature verification

Files created:
- apps/api/src/federation/types/federation-agent.types.ts
- apps/api/src/federation/federation-agent.service.ts
- apps/api/src/federation/federation-agent.service.spec.ts

Files modified:
- apps/api/src/federation/command.service.ts (agent command routing)
- apps/api/src/federation/federation.controller.ts (agent endpoints)
- apps/api/src/federation/federation.module.ts (service registration)
- apps/orchestrator/src/api/agents/agents.controller.ts (status endpoint)
- apps/orchestrator/src/api/agents/agents.module.ts (lifecycle integration)

Testing:
- 12/12 tests passing for FederationAgentService
- All command service tests passing
- TypeScript compilation successful
- Linting passed

Refs #93

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
2026-02-03 14:37:06 -06:00

4.2 KiB
Raw Blame History

Issue ORCH-109: Agent lifecycle management

Objective

Implement agent lifecycle management service to manage state transitions through the agent lifecycle (spawning → running → completed/failed/killed).

Approach

Following TDD principles:

  1. Write failing tests first for all state transition scenarios
  2. Implement minimal code to make tests pass
  3. Refactor while keeping tests green

The service will:

  • Enforce valid state transitions using state machine
  • Persist agent state changes to Valkey
  • Emit pub/sub events on state changes
  • Track agent metadata (startedAt, completedAt, error)
  • Integrate with ValkeyService and AgentSpawnerService

Acceptance Criteria

  • src/spawner/agent-lifecycle.service.ts implemented
  • State transitions: spawning → running → completed/failed/killed
  • State persisted in Valkey
  • Events emitted on state changes (pub/sub)
  • Agent metadata tracked (startedAt, completedAt, error)
  • State machine enforces valid transitions only
  • Comprehensive unit tests with ≥85% coverage
  • Tests follow TDD (written first)

Implementation Details

State Machine

Valid transitions (from state.types.ts):

  • spawning → running, failed, killed
  • running → completed, failed, killed
  • completed → (terminal state)
  • failed → (terminal state)
  • killed → (terminal state)

Key Methods

  1. transitionToRunning(agentId) - Move agent from spawning to running
  2. transitionToCompleted(agentId) - Mark agent as completed
  3. transitionToFailed(agentId, error) - Mark agent as failed with error
  4. transitionToKilled(agentId) - Mark agent as killed
  5. getAgentLifecycleState(agentId) - Get current lifecycle state

Events Emitted

  • agent.running - When transitioning to running
  • agent.completed - When transitioning to completed
  • agent.failed - When transitioning to failed
  • agent.killed - When transitioning to killed

Progress

  • Read issue requirements
  • Create scratchpad
  • Write unit tests (TDD - RED phase)
  • Implement service (TDD - GREEN phase)
  • Refactor and add edge case tests
  • Verify test coverage = 100%
  • Add service to module exports
  • Verify build passes
  • Create Gitea issue
  • Close Gitea issue with completion notes

Testing

Test coverage: 100% (28 tests)

Coverage areas:

  • Valid state transitions (spawning→running→completed)
  • Valid state transitions (spawning→failed, running→failed)
  • Valid state transitions (spawning→killed, running→killed)
  • Invalid state transitions (should throw errors)
  • Event emission on state changes
  • State persistence in Valkey
  • Metadata tracking (timestamps, errors)
  • Conditional timestamp setting (startedAt, completedAt)
  • Agent not found error handling
  • List operations

Notes

  • State transition validation logic already exists in state.types.ts
  • ValkeyService provides state persistence and pub/sub
  • AgentSpawnerService manages agent sessions in memory
  • This service bridges the two by managing lifecycle + persistence

Completion Summary

Successfully implemented ORCH-109 following TDD principles:

Files Created

  1. /home/localadmin/src/mosaic-stack/apps/orchestrator/src/spawner/agent-lifecycle.service.ts - Main service implementation
  2. /home/localadmin/src/mosaic-stack/apps/orchestrator/src/spawner/agent-lifecycle.service.spec.ts - Comprehensive tests (28 tests, 100% coverage)

Files Modified

  1. /home/localadmin/src/mosaic-stack/apps/orchestrator/src/spawner/spawner.module.ts - Added service to module
  2. /home/localadmin/src/mosaic-stack/apps/orchestrator/src/spawner/index.ts - Exported service

Key Features Implemented

  • State transition enforcement via state machine
  • State persistence in Valkey
  • Pub/sub event emission on state changes
  • Metadata tracking (startedAt, completedAt, error)
  • Comprehensive error handling
  • 100% test coverage (28 tests)

Gitea Issue

  • Created: #244
  • Status: Closed
  • URL: #244

Next Steps

This service is now ready for integration with:

  • ORCH-117: Killswitch implementation (depends on this)
  • ORCH-127: E2E test for concurrent agents (depends on this)
Reference in New Issue View Git Blame Copy Permalink