Repository navigation
RUM-18337: Fix slow frame continuity measurement and resend duration corrections in partial view updates - #3980
Conversation
🎉 All green!🧪 All tests passed 🔄 Datadog retried 2 tests - 2 passed on retry 🎯 Code Coverage (details) 🔗 Commit SHA: f4602dd | Docs | View more details | Give us feedback! |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 2b391e8bec
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
![datadog-datadog-prod-us1[bot]](https://avatars.githubusercontent.com/u/365230?s=60&v=4){kind=small}
There was a problem hiding this comment.
Two adjacent 400 ms jank frames now produce one record just below 700 ms, losing over 100 ms from the reported slow-frame timeline. The continuity change exposes the existing duration cap under default configuration.
🤖 Bits Code Review · Commit 2b391e8
408cce1 to
2b391e8
Compare
There was a problem hiding this comment.
More details
Static review supports the revised continuity tracking, duration-cap handling, and resending of duration corrections in partial view updates. No reportable defect was identified.
🤖 Bits Code Review · Commit 4a3ac14
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 4a3ac14006
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
4a3ac14 to
f4602dd
Compare
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: f4602dd3c3
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
aleksandr-gringauz
added
the
ready-for-human-review
3 participants
What does this PR do?
Two related fixes to RUM slow-frame tracking:
DefaultSlowFramesListenernow measures the gap between consecutive slow frames from the end of the previous record (start + duration) instead of its start, so genuinely back-to-back slow frames are correctly merged into a single record instead of incorrectly starting a new one.diffViewEvent'sdiffSlowFramesnow resends the last known slow-frame record in a view update whenever itsdurationhas changed, not just when a newer record with a laterstartappears — fixing a case where a growing record's duration correction was silently dropped from the diff after its first transmission.Motivation
The two bugs compound each other: the first makes slow frames merge far more often once fixed, which increases exposure to the second.
Additional Notes
DefaultSlowFramesListenerTest("separate janks") where the listener was built with a hardcodedcontinuousSlowFrameThresholdNs = 16instead of the test's own forged parameter.Review checklist (to be filled by reviewers)