pr-commenter[bot]
commented
1 month ago Regression Detector (DogStatsD)
Regression Detector Results
Run ID: c8b5a889-66f8-407b-9537-edc6da91f058
Baseline: 7.52.0 Comparison: 7.52.1
Performance changes are noted in the perf column of each table:
- ✅ = significantly better comparison variant performance
- ❌ = significantly worse comparison variant performance
- ➖ = no significant change in performance
No significant changes in experiment optimization goals
Confidence level: 90.00% Effect size tolerance: |Δ mean %| ≥ 5.00%
There were no significant changes in experiment optimization goals at this confidence level and effect size tolerance.
lukesteensen
The goal here is just to start reducing some complexity so that subsequent changes are easier to make and simpler to review. Where previously we had a
Deserializerwrapping a framer wrapping a codec, we now just have a framer and a codec that are used side-by-side. By avoiding the nesting, we're making it easier to introduce a semi-parsed result struct from the codec without needing to thread reference lifetimes through multiple layers of machinery. Actually implementing that change to the codec's parsing function will be the next step, and from there we will be able to rework the rest of the logic (enrichment, differentiating between message types, forwarding, etc) to be a more straightforward function from that semi-parsed representation to a full event, with fewer intermediate steps and less mutation, as well as our ultimate goal, to easily pass in a string/context interner specific to that socket handler.There's likely to be a few bits that were dropped that need to be added back (as well as some things that ought to be simplified further given the new state), so don't hesitate to point anything out in review 😅