Repository navigation
Extend the publish sweep and attribute its limit to generators or Fanout - #2
Merged
Merged
Conversation
…or Fanout A split run on a 32-vCPU Fanout VM doubled its accepted rate with every doubling of generator processes up to 16 per signal while Fanout used about 17 cores, and every phase was classified saturated-or-generator-limited. The publish sweep now runs 1 to 64 processes per signal. Split runs read the generator host's /proc/stat at each phase boundary and record its average CPU; a trial whose generator host averaged at least 90% of its CPUs is generator-limited, and one that gained under 10% over the previous process count with the generator below that is saturated. Read scheduling stops at the measured phase boundary, before the final CPU read. docs/telemetrygen.md explains the generator's own limits: trace exports share one serialized batch processor whose queue drops when full, while logs and metrics export per worker and wait for each acknowledgement. Claude-Session: https://claude.ai/code/session_01XA52nWyQ3R5f5kK8kjpAQW
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
publishpreset sweeps 1, 2, 4, 8, 16, 32 and 64 generator processes per signal, so a split run can pass the point where Fanout stops scaling./proc/statat each phase boundary) and classify trials asgenerator-limited(generator at or above 90% CPU) orsaturated(under 10% gain over the previous process count) instead of the ambiguoussaturated-or-generator-limited.docs/telemetrygen.mddocuments the generator's own limits, including the trace exporter's serialized batch processor and its dropping queue.Verification
just test(tests and vet), ShellCheck andbash -npass./proc/statarithmetic, the new classification cases, and the scheduling boundary with a delayed CPU read.https://claude.ai/code/session_01XA52nWyQ3R5f5kK8kjpAQW