fix: apply the tracer snapshot/log rate policy to sampling defaults (DEBUG-5730)#2241
Draft
p-datadog wants to merge 2 commits into
Draft
fix: apply the tracer snapshot/log rate policy to sampling defaults (DEBUG-5730)#2241p-datadog wants to merge 2 commits into
p-datadog wants to merge 2 commits into
Conversation
When a log probe carries captureExpressions and no explicit sampling, default its sampling_snapshots_per_second to 1/s (snapshot rate) instead of the 5000/s log default. Evaluating capture expressions is cost-equivalent to a full snapshot; running them under the log rate limit spikes apm-processing CPU (the same class of issue that gated Java <1.64.0 and Python <4.11.0).
Contributor
|
Name the two defaults (SNAPSHOT_SAMPLING_SNAPSHOTS_PER_SECOND = 1, LOG_SAMPLING_SNAPSHOTS_PER_SECOND = 5000) and, absent explicit sampling, default any snapshot-emitting log probe (full captureSnapshot OR capture expressions) to the snapshot rate, matching the Ruby tracer's (@capture_snapshot || capture_expressions present) ? 1 : 5000 policy. Plain log probes keep the log rate. Also removes the duplicated 5000 literal on the service-config default.
p-datadog
changed the title
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Motivation
Log probes default
sampling_snapshots_per_secondto5000(the log rate) when no explicit sampling is set. That's correct for plain log probes, but a probe that emits a snapshot — a fullcaptureSnapshotor capture expressions — is cost-equivalent to a snapshot and should be limited like one. Running snapshots under the log rate limit is the DEBUG-5730 apm-processing CPU spike.The Ruby tracer already encodes this policy (
probe.rb):(@capture_snapshot || capture_expressions present) ? 1 : 5000. This applies the same policy in the shared parser, which is where PHP (and other libdatadog consumers) get their defaults.Change
SNAPSHOT_SAMPLING_SNAPSHOTS_PER_SECOND = 1,LOG_SAMPLING_SNAPSHOTS_PER_SECOND = 5000(previously bare literals, and5000was duplicated).Behavior change / open questions
captureSnapshotprobes with no explicit sampling (5000 → 1). That matches Ruby and is arguably a latent-bug fix, but it's a behavior change for any libdatadog consumer relying on the old 5000 default for snapshot probes — please sanity-check.1is imported from the Ruby tracer's convention; confirming it's the intended snapshot rate for the other consumers would be good.Cargo.lockis v4, toolchain is 1.75); relying on CI.