You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
test: measure Keepalive handoffs and verified visitor loads - #811
Extend the opt-in browser handoff harness on top of #810. It reports three different observations: a browser published, its exact running publisher served a checked asset, and the public proxy served a checked page and script. All measurements share one run-local timeline.
Why
A gossip count or successful startup is not evidence that a visitor got the right bytes. We need the time from invitation to independently verified service, the cost of copying the release, and visitor failures during replacement before deciding on a fleet size. None of these measurements needs a production host counter.
Changes
Record tab-local turn milestones and use the exact publisher address, run ID and release hash for independent probes. Chrome reports encoded HTTP bytes separately from decoded bundle size and incomplete transfers.
Sample public visitor loads on a fixed cadence. Report integrity failures, skipped samples, and latency percentiles; keep the deliberate final-host outage separate.
Optionally sample named infrastructure queries on the same timeline. Missing measurements stay null; reports omit backend labels, raw errors and host credentials.
Add a loopback Chrome fixture and instructions for capturing a report on an empty test deployment.
Local verification at 6024e5e9f: mise exec -- task js:keepalive:test (30 tests), mise exec -- task js:keepalive:measurement-check (Chrome with a loopback fixture), mise exec -- task js:keepalive:build, demo TypeScript checks, repository JS lint, formatting and git diff --check passed. The Chrome fixture independently detected modified bytes and HTTP 503 and captured actual encoded HTTP transfers on an expired synthetic turn. The real-handoff Task was dry-run only to check its command line. No hosted run, scale/fault qualification, merge, or deployment was performed.
This branch has not been deployed
No deployments
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
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.
What
Extend the opt-in browser handoff harness on top of #810. It reports three different observations: a browser published, its exact running publisher served a checked asset, and the public proxy served a checked page and script. All measurements share one run-local timeline.
Why
A gossip count or successful startup is not evidence that a visitor got the right bytes. We need the time from invitation to independently verified service, the cost of copying the release, and visitor failures during replacement before deciding on a fleet size. None of these measurements needs a production host counter.
Changes