feat: replace REST API with gRPC service
Complete migration from FastAPI/HTTP to gRPC: Proto definition (proto/daena/v1/daena.proto): - Daena service with 9 RPCs: GetStatus, GetHealth, GetConfig, ListPlugins, GetQueueState, ListPendingRecords, ListDeadRecords, ListSinks, ReloadConfig - Messages for all request/response types with proper field types gRPC server (src/daena/api/grpc_server.py): - Async gRPC aio server with DaenaServicer implementation - DaenaGRPCServer with start/stop/serve_forever lifecycle - All endpoints delegate to DaenaRuntime gRPC client (src/daena/cli/main.py): - CLI now uses grpc.insecure_channel + DaenaStub - Same commands: status, health, queue, sinks, plugins, config, version Config changes: - Remove FastAPI/uvicorn/http config, replace with grpc config - GRPCConfig with host/port/max_workers Dependency changes: - Remove fastapi, uvicorn, starlette, anyio - Add grpcio, grpcio-tools, protobuf - Keep httpx for HTTP sink forwarding (separate from management API) Tests: - 8 new gRPC integration tests (test_grpc.py) - All 31 tests passing, ruff clean
This commit is contained in:
@@ -9,7 +9,7 @@ def test_minimal_config_has_defaults() -> None:
|
||||
config = DaenaConfig()
|
||||
assert config.data_dir == "/var/lib/daena"
|
||||
assert config.logging.level == "info"
|
||||
assert config.api.http.port == 8642
|
||||
assert config.api.grpc.port == 8642
|
||||
assert config.pipeline.batch_size == 100
|
||||
|
||||
|
||||
|
||||
Reference in New Issue
Block a user