90 lines
4.3 KiB
Markdown
90 lines
4.3 KiB
Markdown
# Implementation Plan
|
|
|
|
## Phase 1: Database & Domain Setup
|
|
|
|
- [ ] Create `BacktestJob` entity in `Managing.Domain/Backtests/`
|
|
- [ ] Create `BacktestJobStatus` enum (Pending, Running, Completed, Failed)
|
|
- [ ] Create database migration for `BacktestJobs` table
|
|
- [ ] Add indexes: `idx_status_priority`, `idx_bundle_request`, `idx_assigned_worker`
|
|
- [ ] Create `IBacktestJobRepository` interface
|
|
- [ ] Implement `BacktestJobRepository` with advisory lock support
|
|
|
|
## Phase 2: Compute Worker Project
|
|
|
|
- [ ] Refactor `Managing.Workers.Api` project (or rename to `Managing.Compute`)
|
|
- [ ] Remove Orleans dependencies completely
|
|
- [ ] Add project references to shared projects (Application, Domain, Infrastructure)
|
|
- [ ] Configure DI container with all required services (NO Orleans)
|
|
- [ ] Create `BacktestComputeWorker` background service
|
|
- [ ] Implement job polling logic (every 5 seconds)
|
|
- [ ] Implement job claiming with PostgreSQL advisory locks
|
|
- [ ] Implement semaphore-based concurrency control
|
|
- [ ] Implement progress callback mechanism
|
|
- [ ] Implement heartbeat mechanism (every 30 seconds)
|
|
- [ ] Add configuration: `MaxConcurrentBacktests`, `JobPollIntervalSeconds`, `WorkerId`
|
|
|
|
## Phase 3: API Server Updates
|
|
|
|
- [ ] Update `BacktestController` to create jobs instead of calling grains directly
|
|
- [ ] Implement `CreateBundleBacktest` endpoint (returns immediately)
|
|
- [ ] Implement `GetJobStatus` endpoint (polls database for single job)
|
|
- [ ] Implement `GetBundleStatus` endpoint (polls database, aggregates job statuses)
|
|
- [ ] Update `Backtester.cs` to generate `BacktestJob` entities from bundle variants
|
|
- [ ] Remove all Orleans grain calls for backtests (direct replacement, no feature flags)
|
|
- [ ] Remove `IGrainFactory` dependency from `Backtester.cs`
|
|
|
|
## Phase 4: Shared Logic Extraction
|
|
|
|
- [ ] Create `BacktestExecutor.cs` service (new file)
|
|
- [ ] Extract backtest execution logic from `BacktestTradingBotGrain` to `BacktestExecutor`
|
|
- [ ] Make backtest logic Orleans-agnostic (no grain dependencies)
|
|
- [ ] Add progress callback support to execution method
|
|
- [ ] Ensure candle loading works in compute worker context
|
|
- [ ] Handle credit debiting/refunding in executor
|
|
- [ ] Handle user context resolution in executor
|
|
|
|
## Phase 5: Monitoring & Health Checks
|
|
|
|
- [ ] Add health check endpoint to compute worker (`/health` or `/healthz`)
|
|
- [ ] Add metrics: pending jobs, running jobs, completed/failed counts
|
|
- [ ] Add stale job detection (reclaim jobs from dead workers, LastHeartbeat > 5 min)
|
|
- [ ] Add comprehensive logging for job lifecycle events
|
|
- [ ] Include structured logging: JobId, BundleRequestId, UserId, WorkerId, Duration
|
|
|
|
## Phase 6: SignalR & Notifications
|
|
|
|
- [ ] Inject `IHubContext<BacktestHub>` into compute worker or executor
|
|
- [ ] Send SignalR progress updates during job execution
|
|
- [ ] Update `BacktestJob.ProgressPercentage` in database
|
|
- [ ] Update `BundleBacktestRequest` progress when jobs complete
|
|
- [ ] Send completion notifications via SignalR and Telegram
|
|
|
|
## Phase 7: Deployment
|
|
|
|
- [ ] Create Dockerfile for `Managing.Compute` (or update existing)
|
|
- [ ] Update `docker-compose.yml` to add compute worker service
|
|
- [ ] Configure environment variables: `MaxConcurrentBacktests`, `JobPollIntervalSeconds`, `WorkerId`
|
|
- [ ] Set up health check configuration in Docker
|
|
- [ ] Configure auto-scaling rules for compute workers (min: 1, max: 10)
|
|
|
|
## Phase 9: Testing & Validation
|
|
|
|
- [ ] Unit tests: BacktestJobRepository (advisory locks, job claiming, stale detection)
|
|
- [ ] Unit tests: BacktestExecutor (core logic, progress callbacks)
|
|
- [ ] Integration tests: Single backtest job processing
|
|
- [ ] Integration tests: Bundle backtest with multiple jobs
|
|
- [ ] Integration tests: Concurrent job processing (multiple workers)
|
|
- [ ] Integration tests: Job recovery after worker failure
|
|
- [ ] Integration tests: Priority queue ordering
|
|
- [ ] Load tests: 100+ concurrent users, 1000+ pending jobs, multiple workers
|
|
|
|
## Phase 8: Cleanup & Removal
|
|
|
|
- [ ] Remove or deprecate `BacktestTradingBotGrain.cs` (no longer used)
|
|
- [ ] Remove or deprecate `BundleBacktestGrain.cs` (replaced by compute workers)
|
|
- [ ] Remove Orleans grain interfaces for backtests (if not used elsewhere)
|
|
- [ ] Update `ApiBootstrap.cs` to remove Orleans backtest grain registrations
|
|
- [ ] Remove Orleans dependencies from `Backtester.cs` (keep for other operations)
|
|
- [ ] Update documentation to reflect new architecture
|
|
|