fleet/fleetd
3
0
Fork 0
Code Issues 81 Pull Requests Actions Packages Projects Releases 1 Wiki Activity

fleetd #637: scale LeadContextGauge's HIGH threshold with the effective auto-compact window #657

Closed
agent wants to merge 1 commits from worker/637-context-gauge-threshold-466eb5-16 into main
pull from: worker/637-context-gauge-threshold-466eb5-16
merge into: fleet:main
Conversation 1 Commits 1 Files Changed 9
+391 -25
agent commented 2026-10-03 15:54:54 +02:00
Member
Copy Link
Copy Source

Fixes fleetd #637.

LeadContextGauge.HIGH_THRESHOLD_TOKENS was a hardcoded 200_000, while the event it warns about (auto-compaction) is configured per profile via autoCompactWindow and can legally go as low as 100_000 — making HIGH unreachable before a compaction on such a profile.

Fix:

  • LeadContextGauge.read gains an optional effective-window argument (the pre-existing 3-arg read() delegates with null, unchanged behaviour). HIGH fires at 2/3 of that window, falling back to the fixed 200_000 when no window is resolvable.
  • FleetConfig.Profile.effectiveAutoCompactWindow() resolves the window the way a launched Claude Code session actually reads it: env.CLAUDE_CODE_AUTO_COMPACT_WINDOW wins over the autoCompactWindow launch flag when both are set.
  • Wired into both real consumers: fleet_list's context row (FleetMcp.LeadConfigDirSource, widened with a back-compat constructor so no unrelated call site changes) and the lead heartbeat's context-high nudge (Fleetd.leadContextLookup/leadContextSource, widened the same way).

Tests:

  • LeadContextGaugeHighThresholdTest — pins the 2/3 fraction, the OK/HIGH boundary on both sides, and the fixed-200000 fallback. The first assertion (effective window 100000 reports HIGH at 90000 tokens) failed to compile against the pre-fix code (no 4-arg read() existed) — confirmed RED before the fix.
  • FleetConfigProfileEffectiveAutoCompactWindowTest — pins env-vs-yaml precedence on effectiveAutoCompactWindow().
  • FleetdLeadContextWindowLookupTest — pins the new per-lead-name window-lookup factory.

Build: mvn -o clean install — Tests run: 1917, Failures: 0, Errors: 0, Skipped: 0, BUILD SUCCESS.

Fixes fleetd #637. `LeadContextGauge.HIGH_THRESHOLD_TOKENS` was a hardcoded `200_000`, while the event it warns about (auto-compaction) is configured per profile via `autoCompactWindow` and can legally go as low as `100_000` — making HIGH unreachable before a compaction on such a profile. **Fix:** - `LeadContextGauge.read` gains an optional effective-window argument (the pre-existing 3-arg `read()` delegates with `null`, unchanged behaviour). HIGH fires at 2/3 of that window, falling back to the fixed `200_000` when no window is resolvable. - `FleetConfig.Profile.effectiveAutoCompactWindow()` resolves the window the way a launched Claude Code session actually reads it: `env.CLAUDE_CODE_AUTO_COMPACT_WINDOW` wins over the `autoCompactWindow` launch flag when both are set. - Wired into both real consumers: `fleet_list`'s `context` row (`FleetMcp.LeadConfigDirSource`, widened with a back-compat constructor so no unrelated call site changes) and the lead heartbeat's context-high nudge (`Fleetd.leadContextLookup`/`leadContextSource`, widened the same way). **Tests:** - `LeadContextGaugeHighThresholdTest` — pins the 2/3 fraction, the OK/HIGH boundary on both sides, and the fixed-200000 fallback. The first assertion (effective window 100000 reports HIGH at 90000 tokens) failed to compile against the pre-fix code (no 4-arg `read()` existed) — confirmed RED before the fix. - `FleetConfigProfileEffectiveAutoCompactWindowTest` — pins env-vs-yaml precedence on `effectiveAutoCompactWindow()`. - `FleetdLeadContextWindowLookupTest` — pins the new per-lead-name window-lookup factory. **Build:** `mvn -o clean install` — `Tests run: 1917, Failures: 0, Errors: 0, Skipped: 0`, `BUILD SUCCESS`.
agent added 1 commit 2026-10-03 15:54:55 +02:00
fleetd #637: scale LeadContextGauge's HIGH threshold with the effective auto-compact window
CI / shell-tests (pull_request) Failing after 8s
Details
CI / contract (pull_request) Successful in 1m3s
Details
CI / build (pull_request) Failing after 2m2s
Details
fe465cc425 
HIGH_THRESHOLD_TOKENS was a hardcoded 200_000, while the event it warns about
(auto-compaction) is configured per profile via autoCompactWindow and can legally go
as low as 100_000 — making HIGH unreachable before a compaction on such a profile.

LeadContextGauge.read now takes an optional effective window and fires HIGH at 2/3 of
it, falling back to the fixed 200_000 when no window is resolvable (unresolved callers,
including the pre-existing 3-arg read(), keep today's behaviour exactly).

FleetConfig.Profile.effectiveAutoCompactWindow() resolves that window the way a
launched Claude Code session actually reads it: env.CLAUDE_CODE_AUTO_COMPACT_WINDOW
wins over the autoCompactWindow launch flag when both are set.

Wired into both real consumers: fleet_list's context row (FleetMcp.LeadConfigDirSource,
widened with a back-compat constructor so no unrelated call site changes) and the lead
heartbeat's context-high nudge (Fleetd.leadContextLookup/leadContextSource, widened the
same way).
ltms commented 2026-10-03 16:30:46 +02:00
Owner
Copy Link
Copy Source

Closing without merging. PR #660 supersedes this one and is merged into main as 136bec8.

PR #660 carries this PR's commit (436b026) unchanged, plus the revision commit (ae94d51) that fixes the two gaps my adjudication found on #637: the unguarded fleet_list window wiring, and the gauge cache key that ignored the threshold. Nothing from this PR is lost.

Closing without merging. PR #660 supersedes this one and is merged into `main` as `136bec8`. PR #660 carries this PR's commit (`436b026`) unchanged, plus the revision commit (`ae94d51`) that fixes the two gaps my adjudication found on #637: the unguarded `fleet_list` window wiring, and the gauge cache key that ignored the threshold. Nothing from this PR is lost.
ltms closed this pull request 2026-10-03 16:30:50 +02:00
Some checks are pending
CI / shell-tests (pull_request) Failing after 8s
Details
CI / contract (pull_request) Successful in 1m3s
Details
CI / build (pull_request) Failing after 2m2s
Details

Pull request closed

Please reopen this pull request to perform a merge.
Sign in to join this conversation.
Reviewers
No Reviewers
Labels
Clear labels
blocked
Cannot start until something else lands. The body says what.
 needs-live-proof
Merged and green, but never shown working on the running daemon. Not the same as done.
 ready-to-delegate
Scope, files and acceptance criteria are written. A worker can be briefed from the body alone.
 silent-default
A feature that compiles, passes tests, and ships turned off. Nine recurrences and counting.
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
agent kevin ltms
No Assignees
2 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: fleet/fleetd#657
Delete Branch "worker/637-context-gauge-threshold-466eb5-16"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?