work-unit: control plane — input lifecycle status wiring #1415

Activity

1. strawgate

   added

   enhancementNew feature or requestNew feature or request

   productionRequired for production readinessRequired for production readiness

   P2: mediumMinor bug, missing validation, metric inaccuracyMinor bug, missing validation, metric inaccuracy

   work-unitScheduling ticket for one agent runScheduling ticket for one agent run

   architectureSystemic design improvements that eliminate classes of bugsSystemic design improvements that eliminate classes of bugs

   on Apr 6, 2026
2. github-actions commented on Apr 6, 2026

   

   github-actionsBot

   on Apr 6, 2026 – with GitHub Actions

   Contributor

   More actions

   tl;dr: Outcome A — ready for implementation: input lifecycle state is currently invisible at the shared control-plane boundary, and this can be fixed mechanically by adding per-input lifecycle status to shared diagnostics and wiring explicit transitions in the input loops/receivers.

   Recommendation

   Implement this as a focused wiring slice: add a shared input lifecycle state model, attach it to PipelineMetrics input entries, emit transitions from input poll loops + receiver shutdown/error paths, and expose state in diagnostics JSON while preserving the current quiet-source readiness rule.

   Findings

   Core gaps (verified in code):

   • crates/logfwd-io/src/input.rs:43 (pub trait InputSource: Send) has only poll, name, checkpoint_data, set_offset_by_source; there is no lifecycle/status API.
   • crates/logfwd-io/src/diagnostics.rs:39 stores pipeline inputs as Vec<(String, String, Arc(ComponentStats))> (name/type/stats only).
   • crates/logfwd-io/src/diagnostics.rs:545-555 readiness is let ready = !self.pipelines.is_empty(); (no per-input health/lifecycle influence).
   • crates/logfwd-io/src/diagnostics.rs:1650-1655 test explicitly asserts quiet/no-data pipeline remains ready.
   • crates/logfwd/src/pipeline.rs:1059-1065 poll errors are logged/slept (input.poll_error) but no lifecycle transition is published.
   • crates/logfwd-io/src/otlp_receiver.rs:279-289 and crates/logfwd-io/src/arrow_ipc_receiver.rs:239-244 return empty on no data; no shared lifecycle state updates are emitted.
   • crates/logfwd-io/src/tail.rs:759-766 emits EndOfFile (idle-like signal) but this is not surfaced into diagnostics lifecycle state.

   Related architecture issues (context):

   • meta: I/O control-plane substrate — lifecycle, health, readiness, and status #1306 (shared lifecycle/health/readiness/status substrate)
   • File tailer refactor — split monolith, typestate, per-file correctness #1310 (tailer architecture)
   • OTLP receiver refactor — axum, streaming decompression, unified receivers #1311 (receiver architecture)

   Verification

   $ cargo test -p logfwd-io test_ready_endpoint_ -- --nocapture
   running 2 tests
   test diagnostics::tests::test_ready_endpoint_no_pipelines_returns_503 ... ok
   test diagnostics::tests::test_ready_endpoint_with_pipeline_returns_200 ... ok
   
   test result: ok. 2 passed; 0 failed; 0 ignored; 0 measured; 210 filtered out

   This confirms the current readiness contract: pipeline existence gates readiness, not data freshness.

   Detailed Action Plan (mechanical)

   Files to modify (with current anchors)

   1. crates/logfwd-types/src/diagnostics.rs:10 — pub struct ComponentStats
   2. crates/logfwd-io/src/diagnostics.rs:39 — pub inputs: Vec<(String, String, Arc(ComponentStats))>
   3. crates/logfwd/src/pipeline.rs:1316 — fn build_input_state(...) -> Result(InputState, String)
   4. crates/logfwd/src/pipeline.rs:1038 and :1195 — both input poll loops
   5. crates/logfwd-io/src/otlp_receiver.rs:278 — impl InputSource for OtlpReceiverInput
   6. crates/logfwd-io/src/arrow_ipc_receiver.rs:32 — pub struct ArrowIpcReceiver / receiver lifecycle hooks
   7. crates/logfwd-io/src/input.rs:43 — pub trait InputSource: Send

   Exact changes

   1. Add shared input lifecycle model in logfwd-types

      • In crates/logfwd-types/src/diagnostics.rs, add:
        • enum InputLifecycleState { Starting, Healthy, Idle, Stopping, Stopped, Failed }
        • small lock-free holder/snapshot fields for state, reason (optional), and state_changed_at_ns.
      • Keep API minimal and default-safe (no broad fallbacks).

   2. Attach lifecycle to per-input diagnostics entries

      • In crates/logfwd-io/src/diagnostics.rs, replace input tuple storage with a named struct (e.g. InputMetricsEntry) containing:
        • name, type, stats, lifecycle.
      • Update /api/pipelines JSON serializer to include input lifecycle fields (state, reason, state_changed_at_ns).

   3. Wire lifecycle handle through pipeline input construction

      • In crates/logfwd/src/pipeline.rs:
        • extend InputState to include lifecycle handle.
        • in build_input_state, initialize Starting then set Healthy once source is successfully constructed.

   4. Emit transitions in both input loops (input_poll_loop, input_poll_loop_async)

      • On events.is_empty() following prior healthy activity: set Idle.
      • On data receipt: set Healthy.
      • On poll error branch (input.poll_error): set Failed with reason.
      • On shutdown path before loop break/drain send: set Stopping, then Stopped after drain send attempt.

   5. Receiver-specific stop/failure visibility

      • In crates/logfwd-io/src/otlp_receiver.rs and crates/logfwd-io/src/arrow_ipc_receiver.rs:
        • ensure receiver teardown paths (Drop) can surface Stopping/Stopped through shared lifecycle handle.
        • for disconnected channel conditions in poll, propagate explicit error instead of silently appearing idle.

   6. Preserve quiet-source readiness semantics

      • Keep /ready semantics such that Idle does not fail readiness.
      • readiness should fail only when process/pipeline has fatal lifecycle states per current scope.

   Test to write/update

   1. crates/logfwd-io/src/diagnostics.rs tests:

      • extend test_pipelines_endpoint style assertions to verify lifecycle fields are present for each input.
      • add coverage for representative states: healthy, idle, stopping, failed.
      • retain existing quiet-source readiness behavior (test_ready_endpoint_with_pipeline_returns_200).

   2. crates/logfwd/src/pipeline.rs tests:

      • add focused unit tests for lifecycle transitions in input loops (data->idle, error->failed, shutdown->stopping/stopped).

   3. Receiver tests:

      • crates/logfwd-io/src/otlp_receiver.rs / arrow_ipc_receiver.rs: assert disconnect/stop transitions are reflected.

   Verification commands

   cargo test -p logfwd-io
   cargo test -p logfwd --lib pipeline

   Related Items

   Type	Link	Relevance
   Issue	#1415	This work-unit
   Issue	#1306	Shared lifecycle/readiness/status substrate this wiring plugs into
   Issue	#1310	Tailer lifecycle architecture source
   Issue	#1311	Receiver lifecycle architecture source
   File	crates/logfwd-io/src/input.rs:43	Input boundary lacks lifecycle/status surface
   File	crates/logfwd-io/src/diagnostics.rs:39	Inputs stored as counters-only tuples
   File	crates/logfwd/src/pipeline.rs:1059	Poll errors not wired into lifecycle state
   PR (open)	#1378	Overlaps receiver shutdown/lifecycle paths
   PR (open)	#1402	Overlaps pipeline.rs control flow
   PR (open, draft)	#1370	Touches nearby input/framing/pipeline surfaces

   Note

   🔒 Integrity filtering filtered 10 items

   Integrity filtering activated and filtered the following items during workflow execution.
   This happens when a tool call accesses a resource that does not meet the required integrity or secrecy level of the workflow.

   • issue:work-unit: control plane — input lifecycle status wiring #1415 (issue_read: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".)
   • issue:fix: improve file tailing reliability #805 (issue_read: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".)
   • resource:search_pull_requests (search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".)
   • issue:OTLP receiver refactor — axum, streaming decompression, unified receivers #1311 (issue_read: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".)
   • issue:File tailer refactor — split monolith, typestate, per-file correctness #1310 (issue_read: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".)
   • issue:meta: I/O control-plane substrate — lifecycle, health, readiness, and status #1306 (issue_read: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".)
   • pr:Fix OTLP receiver thread leak and graceful shutdown mechanism #1378 (pull_request_read: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".)
   • pr:fix: resolve config validation regressions (ArrowIPC, Hostnames, Generators) #1381 (pull_request_read: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".)
   • pr:feat(pipeline): evolve Processor trait to chain with error handling, flush, and SmallVec output #1402 (pull_request_read: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".)
   • pr:DO NOT MERGE: Prototype source metadata injection #1370 (pull_request_read: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".)

   ---

   What is this? | From workflow: Issue Triage

   Give us feedback! React with 🚀 if perfect, 👍 if helpful, 👎 if not.
3. strawgate

   added

   component:input/commonCross-cutting input behavior spanning multiple input typesCross-cutting input behavior spanning multiple input types

   component:diagnosticsDiagnostics endpoints, dashboard, health, probes, and trace explorerDiagnostics endpoints, dashboard, health, probes, and trace explorer

   on Apr 6, 2026
4. strawgate

   changed the title [-]work-unit: logfwd-io inputs — lifecycle status wiring for receivers and tailers[/-] [+]work-unit: control plane — input lifecycle status wiring[/+] on Apr 6, 2026
5. coderabbitai

   added

   component:pipelinePipeline orchestration, batching, runtime flow, and worker coordinationPipeline orchestration, batching, runtime flow, and worker coordination

   on Apr 6, 2026
6. strawgate

   closed this as completedin #1491on Apr 7, 2026