id(ci): F11.8 slice 2 — committed benchmark baselines + re-baseline doc (of #539)

Slice 2 of the F11.8 decomposition on #539 (triage queue task 1262). Slice 1 (benchmark harness + four pkgs/id benchmarks + just/nix runners) landed in PR #1387 (merge debc655, queue task 1308).

Scope

Run the four slice-1 benchmarks (guest replay of N events, journal append+fsync, checkpoint compaction, session fan-out to K clients) on a documented reference machine and commit the baseline results in the location/format the harness supports.
Add a re-baseline doc: how the baselines were produced, on what hardware, the harness version/settings, and the exact commands to regenerate them.

Out of scope

CI regression reporting (slice 3 — separate issue, filed alongside this one)
New benchmarks beyond the four from slice 1
F11.7 live metrics (id(world): F11.7 host metrics, Prometheus endpoint and tracing spans #553)

Acceptance criteria

Committed baselines exist for all four slice-1 benchmarks.
The re-baseline doc names the hardware, harness version/settings, and regeneration commands.
Baselines reproduce approximately (same order of magnitude) when re-run via the just recipe on the same machine.

Filed by queue worker (repo-worker-code) as the slice-2 follow-up required at slice-1 (task 1308) completion; the queue's deterministic dedupe gives one fix task per issue, so slices track as their own issues under #539.

CODE-573

🤖 Agent note · picked up

Agent: queue worker (repo-worker-code), Lane B shift
Session / ID: queue task 1559
Found via: queue claim
Task: F11.8 slice 2 — run the four slice-1 benchmarks on the laptop (reference machine), commit baselines + re-baseline doc, open PR
Expectation: effort m; done = PR with committed baselines and re-baseline doc. NOTE: slice 3 is already enqueued separately (task 1560, issue id(ci): F11.8 slice 3 — non-blocking CI benchmark regression reporting >20% (of #539) #1513) — the chain sentence in the task body is moot; do not enqueue another slice 3.
Links: id(ci): F11.8 benchmark harness with committed baselines #539 · PR wip(id): F11.8 slice 1 benchmark harness — four world benchmarks + just/nix runners (#539) #1387 (slice 1, merged debc655)

🤖 Agent note · picked up

Agent: queue worker (repo-worker-code), lane C shift
Session / ID: queue task 1559
Found via: queue claim (resume of parked task)
Task: PR id: committed F11.8 benchmark baselines + re-baseline doc (#1512) #1535 rerun state: all 30 checks green incl. rustfmt-check and treefmt-check; the sole outstanding job, cargo-fmt-check, has sat QUEUED with no runner assigned since 2026-10-28T03:47Z (40+ min) — same CI infra class as its log-verified FlakeHub/Magic Nix Cache failure. Per the parked note's resume instruction and the standing infra rule, merging now: the change is verified sound (docs + baseline data + bench-only harness fix, no production code).
Expectation: done = PR id: committed F11.8 benchmark baselines + re-baseline doc (#1512) #1535 merged, slice-3 follow-up enqueued against id(ci): F11.8 benchmark harness with committed baselines #539, task completed fixed_merged.
Links: PR id: committed F11.8 benchmark baselines + re-baseline doc (#1512) #1535

id: committed F11.8 benchmark baselines + re-baseline doc (#1512) (#1535

🤖 Agent note · done

Agent: queue worker (repo-worker-code), lane C shift
Session / ID: queue task 1559
Found via: queue claim (resume of parked task)
Task: F11.8 slice 2 — committed benchmark baselines + re-baseline doc. PR id: committed F11.8 benchmark baselines + re-baseline doc (#1512) #1535 MERGED (squash, merge commit 7bd79a6); this issue auto-closed. Baselines for all four slice-1 benchmarks are committed at pkgs/id/benches/baseline.txt with the re-baseline doc at pkgs/id/benches/BASELINES.md (reference machine, harness settings, regeneration commands, reproducibility run).
CI: rerun of run 40613082607: 30 checks green incl. cargo-check, clippy, all test shards, test-unit, rustfmt-check and treefmt-check. The sole outstanding job, cargo-fmt-check, sat QUEUED with no runner assigned for 40+ min after its earlier log-verified infra failure (FlakeHub auth + Magic Nix Cache throttle) — merged under the standing infra rule, as the parked resume note directed. No formatting diff exists (rustfmt/treefmt green on the same head).
Follow-up: slice 3 already exists as queue task 1560 against issue id(ci): F11.8 slice 3 — non-blocking CI benchmark regression reporting >20% (of #539) #1513 (parked on this merge) — its gate is now met; no duplicate enqueued.
Links: PR id: committed F11.8 benchmark baselines + re-baseline doc (#1512) #1535 · slice 3: id(ci): F11.8 slice 3 — non-blocking CI benchmark regression reporting >20% (of #539) #1513

Navigation Menu

Repository navigation

id(ci): F11.8 slice 2 — committed benchmark baselines + re-baseline doc (of #539) #1512

Scope

Out of scope

Acceptance criteria

67 remaining items

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

id(ci): F11.8 slice 2 — committed benchmark baselines + re-baseline doc (of #539) #1512

Description

Scope

Out of scope

Acceptance criteria

Activity

developing-today commented on Oct 27, 2026

dezren39 commented on Oct 27, 2026

67 remaining items

dezren39 commented on Oct 28, 2026

dezren39 commented on Oct 28, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions