blog: one more reading, then it cleared - #931
Open
gHashTag wants to merge 1 commit into
Open
Conversation
Adds a bilingual post (EN + RU) about the disk-halt hysteresis mechanism built for the trinity-fpga autonomous loop's tripwire system, and its first real-world test: a genuine disk crisis (0.18 GiB -> 127 MiB free) that the loop correctly held at HALT for one extra confirming reading before clearing, rather than flapping to RUNNING on the first good reading after recovery. Receipts point at the three relevant trinity-fpga commits (the original tripwires, the hysteresis/flap-detection addition, and the commit recording the real crisis). Open questions are explicit about what this does and does not show: the state machine is verified against production data, but the crisis's root cause was never identified, and flap detection itself has not yet fired for real in this run. Verified with npm run build:ci (exit 0) and by grepping the built per-slug chunk for both English and Russian body content.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
A bilingual (EN + RU) blog post about the trinity-fpga autonomous loop's disk-halt hysteresis mechanism, and its first real-world test.
The story: a tripwire that only looks at the current disk reading has no memory — a bounce near a threshold flips the verdict every cycle. Built two small state machines (hysteresis requiring 2 consecutive recovery readings before clearing; flap detection for repeated halt episodes) three days before a genuine disk crisis (0.18 GiB -> 127 MiB free) tested them for real. The mechanism held the verdict at HALT for one extra confirming reading even though the raw disk state had already recovered, then cleared on the next check — exactly the designed behavior, now proven against production data instead of only synthetic test fixtures.
Receipts: three trinity-fpga commits (the original tripwire mechanism, the hysteresis/flap-detection addition, and the commit recording the real crisis + proof).
Open questions are explicit about scope: the crisis's root cause was never identified (something outside the loop's visibility resolved it); flap detection hasn't fired for real yet (only tested against seeded synthetic data); a detected flap currently only prints a warning rather than writing a formal anomaly record.
Test plan
npm run build:ciexits 0one-more-reading-then-it-cleared-*.js) greped for both EN and RU body content — confirmed present🤖 Generated with Claude Code