feat(observability): add LFS phase metrics and snapshot serve bandwidth histogram#318
Merged
Merged
Conversation
worstell
changed the title
worstell
force-pushed
the
feat/lfs-and-bandwidth-metrics
branch
from
078b88e to
c8fc785
Compare
![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: c8fc7857a1
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".
inez
approved these changes
May 19, 2026
worstell
force-pushed
the
feat/lfs-and-bandwidth-metrics
branch
from
c8fc785 to
265fee5
Compare
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 265fee5f2f
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".
…th histogram Adds two pieces of visibility motivated by staging benchmark findings: 1. cachew.git.snapshot_serve_bandwidth_mbps (per-request MiB/s, by source and repository). Aggregate bytes/duration averages are ambiguous — e.g. cash-server cached serves averaged ~325 MiB/s while a raw curl from the same workstation saw ~588 MiB/s. A per-request distribution lets us see whether slow clients pull the tail down vs the server itself being slow. 2. cachew.git.lfs_phase_duration_seconds and cachew.git.lfs_phase_bytes for LFS-snapshot generation phases (discover, clone, fetch, archive_upload). LFS snapshot generation is the biggest server-side cost in staging (~8.7 min average), and today we only see total duration; per-phase breakdown is needed to know whether to target clone, LFS fetch, or pack/upload. Also stamps cachew.snapshot.bandwidth_mbps onto the active snapshot span so trace samples carry the same value. Co-authored-by: Amp <amp@ampcode.com> Amp-Thread-ID: https://ampcode.com/threads/T-019e41af-0a15-718d-a9d8-e26df6071f9b
worstell
force-pushed
the
feat/lfs-and-bandwidth-metrics
branch
from
265fee5 to
77b4ac2
Compare
worstell
added a commit
that referenced
this pull request
May 19, 2026
The bandwidth histogram added in #318 topped out at 5000 MiB/s (~5.2 GB/s), so any serve from ~5 GiB/s through the cachew server NIC ceiling collapses into the +Inf bucket and we lose the signal where we most need it. Add 10000 and 15000 MiB/s buckets so cachew.git.snapshot_serve_bandwidth_mbps can distinguish 'saturating a 10 GbE workstation' from 'approaching the server NIC limit', with some headroom past the theoretical max to spot misattribution. Amp-Thread-ID: https://ampcode.com/threads/T-019e41af-0a15-718d-a9d8-e26df6071f9b Co-authored-by: Amp <amp@ampcode.com>
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
2 participants
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.
Adds two pieces of visibility motivated by staging benchmark findings:
cachew.git.snapshot_serve_bandwidth_mbps— per-request MiB/s, keyed by source and repository. Aggregate bytes/duration averages are ambiguous:cash-servercached serves averaged ~325 MiB/s while a rawcurlfrom the same workstation saw ~588 MiB/s. A per-request distribution lets us tell whether slow clients pull the tail down vs the server itself being slow. Also stamped onto the active snapshot span ascachew.snapshot.bandwidth_mbps.cachew.git.lfs_phase_duration_secondsandcachew.git.lfs_phase_bytes— broken out by phase (discover,clone,fetch,archive_upload). LFS-snapshot generation is the biggest server-side cost in staging (~8.7 min average), and today we only see total duration; per-phase breakdown is needed to know whether to target clone, LFS fetch, or pack/upload.Both histograms use explicit buckets sized for the values we expect.