chore: implement phase 9

[Tofel]

Implements Check(), the I/O shell around pure decide() (invariants H5/H7). Single-step mode synthesizes the header and its own sentinel, so no mode flag reaches the pure layer; log mode stops the recorder before the one full read. The header (not a re-resolved definition) decides skipped, the drain set, and transitionGrace; the flock (not the pidfile) is the authority for whether a writer still exists.

Review focus: check.go mode handling, and the header-vs-flock authority decisions.

[github-actions]

📊 API Diff Results

No changes detected for module github.com/smartcontractkit/chainlink-testing-framework/grafana-alertcheck

View full report

[Copilot]

🔵 Needs a closer look

It introduces a large, correctness-critical orchestration layer (check.go) with substantial new behavior despite strong tests, so final human review is warranted.

Pull request overview

Implements Phase 9 by adding the Check() I/O shell around the existing pure classification/coverage logic, with explicit handling for single-step vs recorded-log modes and tightened “authority” decisions (header vs re-resolved definitions; flock vs pidfile) to avoid fail-open behavior.

Changes:

• Added Check() (and helpers) to orchestrate validation, definition resolution, timing derivation, evidence collection, recorder shutdown via flock authority, drain wait, and final decide() classification.
• Centralized first-round polling into firstObservations() for both watch-start and single-step measurement passes, including §3.2 verification and stable poll ordering.
• Updated timing/skipped logic so “paused at window start” is sourced from the log header (or equivalent start-of-step snapshot) rather than post-window re-resolved definitions; added supporting helpers and tests.

File summaries

File	Description
grafana-alertcheck/internal/gate/watch.go	Extracts firstObservations() and uses it for initial heartbeats/latency measurement prior to budget checking.
grafana-alertcheck/internal/gate/watch_daemon_test.go	Adds a lock-holder test helper process to validate flock-based recorder stop behavior.
grafana-alertcheck/internal/gate/schedule.go	Makes transitionGrace exclusion depend on “paused at start” authority (header/snapshot), not post-window definition state.
grafana-alertcheck/internal/gate/schedule_test.go	Pins the new header-authoritative grace exclusion behavior in log mode.
grafana-alertcheck/internal/gate/log.go	Adds Header.pausedAtStart(), stateRuleByUID(), and ReadLogHeader() to support early identity checks and shared UID selection logic.
grafana-alertcheck/internal/gate/flock.go	Adds tryLockExclusive() to distinguish contention from actual flock failures for “writer exists?” probing.
grafana-alertcheck/internal/gate/coverage.go	Updates commentary to reflect Phase 9/P8 responsibilities now implemented elsewhere.
grafana-alertcheck/internal/gate/classify.go	Switches skipped-rule detection to header authority via pausedAtStart.
grafana-alertcheck/internal/gate/classify_test.go	Updates tests to construct pause authority via header rather than Definition.IsPaused.
grafana-alertcheck/internal/gate/check.go	New: full Phase 9 Check() implementation including validation, early header read, stop protocol, drain wait, and result merge.
grafana-alertcheck/internal/gate/check_test.go	New: comprehensive integration-style tests for single-step and recorder modes, including edge cases and regressions.
grafana-alertcheck/internal/gate/check_process.go	New: process signaling helper (signalRecorder) used by recorder stop logic.

Review details

• Files reviewed: 12/12 changed files
• Comments generated: 1
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.