[Serverless] Reset both inferred spans on each invocation. #30881
Conversation
github-actions
bot
added
the
short review
purple4reina
added
changelog/no-changelog
team/serverless
qa/done
Serverless Benchmark Results
tl;drUse these benchmarks as an insight tool during development.
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.
I need more helpFirst off, do not worry if the benchmarks are failing. They are not tests. The intention is for them to be a tool for you to use during development. If you would like a hand interpreting the results come chat with us in 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=48518985 --os-family=ubuntuNote: This applies to commit aa1420c |
Regression DetectorRegression Detector ResultsMetrics dashboard Baseline: 2a4706d Optimization Goals: ❌ Significant changes detected
|
| perf | experiment | goal | Δ mean % | Δ mean % CI | trials | links |
|---|---|---|---|---|---|---|
| ➖ | basic_py_check | % cpu utilization | +2.80 | [-0.98, +6.58] | 1 | Logs |
| ➖ | uds_dogstatsd_to_api_cpu | % cpu utilization | +1.83 | [+1.11, +2.55] | 1 | Logs |
| ➖ | quality_gate_idle_all_features | memory utilization | +0.77 | [+0.66, +0.88] | 1 | Logs bounds checks dashboard |
| ➖ | file_to_blackhole_500ms_latency | egress throughput | +0.13 | [-0.11, +0.36] | 1 | Logs |
| ➖ | file_to_blackhole_100ms_latency | egress throughput | +0.02 | [-0.24, +0.29] | 1 | Logs |
| ➖ | file_to_blackhole_1000ms_latency | egress throughput | +0.02 | [-0.47, +0.50] | 1 | Logs |
| ➖ | tcp_dd_logs_filter_exclude | ingress throughput | -0.00 | [-0.01, +0.01] | 1 | Logs |
| ➖ | file_to_blackhole_0ms_latency | egress throughput | -0.00 | [-0.38, +0.38] | 1 | Logs |
| ➖ | uds_dogstatsd_to_api | ingress throughput | -0.01 | [-0.11, +0.10] | 1 | Logs |
| ➖ | file_to_blackhole_300ms_latency | egress throughput | -0.04 | [-0.23, +0.15] | 1 | Logs |
| ➖ | idle | memory utilization | -0.10 | [-0.15, -0.05] | 1 | Logs bounds checks dashboard |
| ➖ | tcp_syslog_to_blackhole | ingress throughput | -0.28 | [-0.34, -0.22] | 1 | Logs |
| ➖ | file_tree | memory utilization | -0.30 | [-0.43, -0.17] | 1 | Logs |
| ➖ | idle_all_features | memory utilization | -0.33 | [-0.43, -0.23] | 1 | Logs bounds checks dashboard |
| ➖ | quality_gate_idle | memory utilization | -0.71 | [-0.75, -0.66] | 1 | Logs bounds checks dashboard |
| ✅ | pycheck_lots_of_tags | % cpu utilization | -5.20 | [-8.42, -1.99] | 1 | Logs |
Bounds Checks: ❌ Failed
| perf | experiment | bounds_check_name | replicates_passed | links |
|---|---|---|---|---|
| ❌ | idle | memory_usage | 8/10 | bounds checks dashboard |
| ❌ | quality_gate_idle | memory_usage | 8/10 | bounds checks dashboard |
| ❌ | quality_gate_idle_all_features | memory_usage | 9/10 | bounds checks dashboard |
| ✅ | file_to_blackhole_0ms_latency | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_1000ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_100ms_latency | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_100ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_300ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_500ms_latency | memory_usage | 10/10 | |
| ✅ | idle_all_features | memory_usage | 10/10 | bounds checks dashboard |
Explanation
Confidence level: 90.00%
Effect size tolerance: |Δ mean %| ≥ 5.00%
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
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 |
Devflow running:
|
What does this PR do?
When a new invocation starts, the
resourceHandleris reset. In addition to resetting the first inferred span, also reset the second by setting it tonil.Motivation
A single lambda can have multiple event types that invoke it. When some of the invocations would create 2 inferred spans, and some of the invocations would create 1 inferred span, we must ensure that both inferred spans are reset.
This was leading to some really weird behavior when I tried to add sns->sqs trace propagation to the Trace Propagation Functional Tests. I was seeing weird stuff like this, where an
sqs.sendMessage, callsaws.snsinferred span, callsaws.lambda.urlinferred span, callsaws.lambdaspan.Describe how to test/QA your changes
I built an extension layer with this change as
arn:aws:lambda:us-east-1:425362996713:layer:Datadog-Extension-REY:8and ran the Trace Propagation Functional Tests and finally saw them all passing.Possible Drawbacks / Trade-offs
Additional Notes
datadoghq.atlassian.net/browse/SVLS-5732