The fuel floor scales with the lanes - #4327
Merged
Merged
Conversation
A fixed floor is the same mistake the lane count was. 24 was chosen when the swarm ran ten bees; the day it ran twenty, one reading showed about 12 dispatchable issues for 20 lanes while the floor said 24 - barely one round. `--runway auto` is twice the capacity the swarm reports on that same reading, which the feeder already fetches, with a floor of 4 for a capacity that reads zero. A number still means that number and 0 still means off. Measured after the change: `runway: about 12 dispatchable issue(s) against a floor of 40`, with 221 uncovered specs to draw on. Closes #4326 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
gHashTag
enabled auto-merge (squash)
September 20, 2026 12:10
Measured 2026-09-20T12:10Z: the feeder read "about 123 dispatchable issue(s) against a floor of 40" and added nothing, while the last scanning tick had counted 68 claimed and 42 completed - about 12 dispatchable for 20 lanes. A tick that refuses on capacity never scans the board, so its skipSummary is empty, and an empty summary reads exactly like a board with nothing claimed. Same defect, second place: the pusher was given `skips_are_fresh` an hour earlier for the same reason. The claimed and completed counts now come from the pusher's last SCANNED reading when this tick did not scan, and the feeder falls back to the free lanes and says so when no scanned reading is available. Reading the pusher's issue is a coupling and it is named as one in the code: the alternative is a feeder that cannot judge the runway in exactly the state it exists for, because a swarm at full capacity is when its queue empties fastest. Closes #4326 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This was referenced Sep 20, 2026
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
Contributor
Contributor
PR DashboardGenerated at: 2026-09-20 12:30:25 UTC
Summary
Seal Status
|
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
2 participants
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.
Closes #4326
--runway 24was chosen when the swarm ran ten bees. It runs twenty now:A fixed floor is the same mistake the lane count was — a number somebody chose,
which cannot notice that the thing it stands for has changed.
auto= twice the capacity the swarm reports on that same reading, whichthe feeder already fetches. A number still means that number;
0still meansoff.
The 68 claimed are not a bug to fix here: a row that spent its retry ceiling is
released once, after an hour, then stays rejected on purpose — "a third
identical failure is evidence about the issue, not about the attempt".
🤖 Generated with Claude Code