Repository navigation
study: report local outages apart from provider failures - #16
Merged
Merged
Conversation
The collector polls every provider in the same tick, so a client-side "fetch failed" on most of them at once is the collector's own connection dropping. On a 64.6 h laptop run that was 15.7% of ticks, and it showed up as a ~9.5% failure rate on every provider. Those ticks are now counted in their own section and left out of availability, latency and head lag. The "JSON-RPC error inside 200" column also counted errors on any status code, so it always matched transport failures. It now counts 200s only. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The study report blamed the collector's own connection on the providers.
Local outages. All providers are polled in the same tick. When more than half of them fail in one tick with a client-side
fetch failed, that's the collector's network (Wi-Fi drop, laptop sleep), not any provider. Those ticks now get their own section and are excluded from availability, latency and head lag. Detector findings and incidents are left as recorded.Error inside 200. The column counted
error_in_bodyon any status, so it always equalled transport failures. It now counts responses below 400 only.Run against the 64.6 h study (5 public mainnet providers):
Disagreements at equal height are unchanged (they already ignore failed answers): 2 of 7,000
eth_callruns.Tested: ran the new report inside the study container against the live study database. Linted and prettier-clean.
🤖 Generated with Claude Code