[SVLS-4142] Create a cold start span only when a cold start has not already occurred #22202
Conversation
DylanLovesCoffee
added
kind/bug
changelog/no-changelog
[deprecated] qa/skip-qa - use other qa/ labels
DylanLovesCoffee
removed
the
[deprecated] qa/skip-qa - use other qa/ labels
Serverless Benchmark Results
tl;dr
What is this benchmarking?The The benchmark is run using a large variety of lambda request payloads. In the charts below, there is one row for each event payload type. How do I interpret these charts?The charts below comes from The benchstat docs explain how to interpret these charts.
Benchmark stats |
Test changes on VMUse this command from test-infra-definitions to manually test this PR changes on a VM: inv create-vm --pipeline-id=32626796 --os-family=ubuntu |
Regression DetectorRegression Detector ResultsRun ID: 8789a510-6119-455e-b219-065aa793b640 Performance changes are noted in the perf column of each table:
No significant changes in experiment optimization goalsConfidence level: 90.00% There were no significant changes in experiment optimization goals at this confidence level and effect size tolerance.
|
| perf | experiment | goal | Δ mean % | Δ mean % CI |
|---|---|---|---|---|
| ❌ | file_to_blackhole | % cpu utilization | +22.95 | [+17.36, +28.53] |
Fine details of change detection per experiment
| perf | experiment | goal | Δ mean % | Δ mean % CI |
|---|---|---|---|---|
| ❌ | file_to_blackhole | % cpu utilization | +22.95 | [+17.36, +28.53] |
| ➖ | basic_py_check | % cpu utilization | +1.03 | [-1.36, +3.41] |
| ➖ | uds_dogstatsd_to_api_cpu | % cpu utilization | +0.85 | [-2.03, +3.73] |
| ➖ | pycheck_1000_100byte_tags | % cpu utilization | +0.43 | [-3.85, +4.70] |
| ➖ | process_agent_real_time_mode | memory utilization | +0.32 | [+0.27, +0.36] |
| ➖ | file_tree | memory utilization | +0.27 | [+0.18, +0.37] |
| ➖ | tcp_dd_logs_filter_exclude | ingress throughput | +0.01 | [-0.04, +0.06] |
| ➖ | uds_dogstatsd_to_api | ingress throughput | +0.01 | [-0.20, +0.21] |
| ➖ | otel_to_otel_logs | ingress throughput | -0.00 | [-0.38, +0.37] |
| ➖ | trace_agent_msgpack | ingress throughput | -0.00 | [-0.01, +0.01] |
| ➖ | trace_agent_json | ingress throughput | -0.00 | [-0.02, +0.01] |
| ➖ | process_agent_standard_check_with_stats | memory utilization | -0.24 | [-0.30, -0.19] |
| ➖ | idle | memory utilization | -0.43 | [-0.48, -0.38] |
| ➖ | process_agent_standard_check | memory utilization | -0.85 | [-0.91, -0.78] |
| ➖ | tcp_syslog_to_blackhole | ingress throughput | -7.89 | [-28.46, +12.68] |
Explanation
A regression test is an A/B test of target performance in a repeatable rig, where "performance" is measured as "comparison variant minus baseline variant" for an optimization goal (e.g., ingress throughput). Due to intrinsic variability in measuring that goal, we can only estimate its mean value for each experiment; we report uncertainty in that value as a 90.00% confidence interval denoted "Δ mean % CI".
For each experiment, we decide whether a change in performance is a "regression" -- a change worth investigating further -- if all of the following criteria are true:
-
Its estimated |Δ mean %| ≥ 5.00%, indicating the change is big enough to merit a closer look.
-
Its 90.00% confidence interval "Δ mean % CI" does not contain zero, indicating that if our statistical model is accurate, there is at least a 90.00% chance there is a difference in performance between baseline and comparison variants.
-
Its configuration does not mark it "erratic".
|
/merge |
|
🚂 MergeQueue Pull request added to the queue. This build is next! (estimated merge in less than 49m) Use |
…lready occurred (#22202) * create a Lambda span on timeouts * don't create a cold start span when the runtime restarts during timeouts * fix linting * fix test * lint: rename name variables * lint again * small fixes * refactor timeout span logic * add mutexes * fix span completed check * revert refactor * remove cold start span changes * use mutex over rwmutex * test routes * add comment + update tests * test endExecutionSpan * add serverless.go test * add test /hello for route * prevent additional cold start spans from being created after runtime is killed * fix bad conflict merge * fix test
What does this PR do?
Checks whether a cold start request ID has been recorded (from a restored execution context) prior to creating a cold start span. Typically this should not be needed since we use
sync.Onceto process the span just once, but after a runtime is killed in the Lambda environment during theINVOKEphase, on the next invocation the ColdStartSpanCreator is reinitialized and thesync.Oncecounter is reset to a value ofdone: 0rather thandone: 1, resulting in another cold start span being created. This will always happen on an invocation that follows a killed runtime.This is noticeable if a cold start invocation hits an out of memory or timeout error. The subsequent invocation will have a trace with another cold start span. The work being done to introduce the
Impending Timeoutspan also makes this noticeable:coldstartspan.mov
Motivation
An argument can be made that after the runtime is killed from a time out or out of memory error, the next invocation must re-initialize the Lambda environment again so it is a cold start. This specific scenario appears to be classified as a suppressed init: https://docs.aws.amazon.com/lambda/latest/dg/telemetry-schema-reference.html#InitPhase.
Additional Notes
Possible Drawbacks / Trade-offs
Describe how to test/QA your changes
Reviewer's Checklist
Triagemilestone is set.major_changelabel if your change either has a major impact on the code base, is impacting multiple teams or is changing important well-established internals of the Agent. This label will be use during QA to make sure each team pay extra attention to the changed behavior. For any customer facing change use a releasenote.changelog/no-changeloglabel has been applied.qa/skip-qalabel, with required eitherqa/doneorqa/no-code-changelabels, are applied.team/..label has been applied, indicating the team(s) that should QA this change.need-change/operatorandneed-change/helmlabels have been applied.k8s/<min-version>label, indicating the lowest Kubernetes version compatible with this feature.