Add additional metrics for remote byte stores #20138
Merged
Conversation
This file contains 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
tgolsson
approved these changes
Nov 2, 2023
src/rust/engine/remote_provider/remote_provider_opendal/src/lib.rs
Outdated
Show resolved
Hide resolved
| } | ||
| Err(_) => Metric::RemoteStoreWriteErrors, | ||
| }; | ||
| workunit.increment_counter(result_metric, 1); |
There was a problem hiding this comment.
Merging this single step into the two branches above would make the code flow easier, I think. And same for next file.
There was a problem hiding this comment.
As in, have two calls to increment_counter?
It might read a bit better, but my thinking was that it's nice to "guarantee" that we record the result against some metric, on every code path, which this style does better than having to remember to call increment_counter on every branch. Only a minor objection, though.
huonw
added a commit
that referenced
this pull request
Nov 5, 2023
Prompted by #20138 (comment), this adds two helpers for a common pattern: increment a metric counter or record a metric observation _if_ the current thread has a workunit handle set. As a related drive-by, this also notices that `increment_counter` takes `&mut self` but is happy with `&self` (and similarly for one `record_observation` function), and so swaps it to use `&self`.
|
I'm going to merge this, @stuhood please let me know if I've missed some context about these metrics and I can fix it up retrospectively. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
This adds a bunch of additional metrics to provide more insight into the byte store behaviour: for each of the "read", "write" and "list missing digests" (aka "exists") operations, track:
This mimics the metrics for the remote action cache.
The minor shuffling of code in this PR also fixes a bug with the
RemoteStoreBlobBytesDownloadedmetric: before this PR, any successful network request would count as bytes downloaded, even if the digest didn't exist/wasn't downloaded (i.e. it successfully determined that the digest wasn't cached). This PR restricts the metric to only count digest bytes when the network request is successful and the digest actually existed/was downloaded.I'm hopeful this will at least give more insight into the scale of the problem in #20133, in addition to being generally useful as "what's pants up to" reporting.