Skip to content

fix(analyze): treat a load-generator burst as an invalid step - #13

Merged
psanders merged 1 commit into
mainfrom
fix/loadgen-burst
Nov 4, 2026
Merged

psanders merged 1 commit into
mainfrom
fix/loadgen-burst

Conversation

@psanders

@psanders psanders commented Nov 4, 2026

Copy link
Copy Markdown
Member

A SIPp that falls behind its schedule later sends the calls it owes in one burst. On Routr's 2026-11-04 calibration that burst (29,674 calls in 10 s at a 2,100/s target) failed the Go EdgePort's 2,100 step and was published as its limit, though the EdgePort handled 5,000/s with no retransmission in a lab run on the same droplet size. A step that placed more than 105% of the offered rate is now invalid, so the verdict is inconclusive instead of a false limit.

Verify: npm test (new case: "should be inconclusive when SIPp placed 110% of the offered rate").

🤖 Generated with Claude Code

A SIPp that falls behind its schedule later sends the calls it owes in one burst. On Routr's
2026-11-04 calibration that burst (29,674 calls in 10 s at a 2,100/s target) made the Go
EdgePort's 2,100 step fail and was published as its limit, though the EdgePort handled
5,000/s without a retransmission in a lab run. A step that placed more than 105% of the
offered rate is now invalid, so the verdict is inconclusive instead of a false limit.
@psanders
psanders merged commit fc06c05 into main Nov 4, 2026
2 checks passed
@psanders
psanders deleted the fix/loadgen-burst branch November 5, 2026 14:26
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant