docs(benchmarks): stop instructing a rebuild of an instrument that exists - #42
Merged
Conversation
…ists THROUGHPUT-STATUS-2026-07-10 tells the reader four times that the per-process CPU-attribution instrument "DOES NOT EXIST YET -- it must be BUILT" and calls building it the next piece of work. PLAN-ENGINE-ATTRIBUTION's backlog actions still list closing #208 and holding #220 at P3 as pending. All of that is stale. #220 shipped: ProcSample.cpu_pids records the exact PID set summed each tick (harness/load/connscale/probe.py:57-70,261), the piecewise max(0, delta) fold is in connscale/runner.py, and falsifiers live in tests/test_connscale_cpu_probe.py. #208 is closed, and its only residual is off-repo measurement that no in-repo change can close. An instruction to build something that already exists is precisely how merged work gets rebuilt -- the same failure the backlog reconcile just addressed at the ledger level. These two documents were the remaining copy of it outside docs/BACKLOG.md. Corrected by annotation, not rewriting. Both files are DATED artifacts: a status record and a plan. Their findings and reasoning were sound for what was known at the time, and silently editing them would falsify the account of why the decisions were made. So the original text stands, with a correction banner at the top of each and an inline supersession marker at the four imperative "build it" sites -- a reader landing mid-document still gets the correction. Verified: both relative links resolve from the docs/benchmarks/ directory; leak gate exit 0 with detectors live (names=7, estate=13, site_prefixes=1).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
THROUGHPUT-STATUS-2026-07-10tells the reader four times that the per-process CPU-attribution instrument "DOES NOT EXIST YET — it must be BUILT", andPLAN-ENGINE-ATTRIBUTION's backlog actions still list closing #208 and holding #220 at P3 as pending work.All of that is stale. #220 shipped:
ProcSample.cpu_pidsrecords the exact PID set summed each tick (probe.py:57-70,261), the piecewisemax(0, Δcpu)fold is inconnscale/runner.py, and falsifiers live intests/test_connscale_cpu_probe.py. #208 is closed, and its only residual is off-repo measurement that no in-repo change can close.An instruction to build something that already exists is precisely how merged work gets rebuilt — the same failure #41 just fixed at the ledger level. These two documents were the remaining copy of it outside
docs/BACKLOG.md.Corrected by annotation, not rewriting
Both files are dated artifacts — a status record and a plan. Their findings and reasoning were sound for what was known at the time, and silently editing them would falsify the account of why those decisions were made. So the original text stands, with a correction banner at the top of each and an inline supersession marker at the four imperative "build it" sites, so a reader landing mid-document still gets the correction.
Verified: both relative links resolve from
docs/benchmarks/; leak gate exit 0 with detectors live (names=7, estate=13, site_prefixes=1).