[serverless] S3 Downstream Span Pointers #32060
Conversation
Test changes on VMUse this command from test-infra-definitions to manually test this PR changes on a VM: inv aws.create-vm --pipeline-id=51115769 --os-family=ubuntuNote: This applies to commit f804396 |
nhulston
added
the
qa/done
Package size comparisonComparison with ancestor Diff per package
Decision✅ Passed |
Regression DetectorRegression Detector ResultsMetrics dashboard Baseline: c50e6cf Optimization Goals: ✅ No significant changes detected
|
| perf | experiment | goal | Δ mean % | Δ mean % CI | trials | links |
|---|---|---|---|---|---|---|
| ➖ | quality_gate_idle_all_features | memory utilization | +1.69 | [+1.58, +1.79] | 1 | Logs bounds checks dashboard |
| ➖ | otel_to_otel_logs | ingress throughput | +0.96 | [+0.22, +1.69] | 1 | Logs |
| ➖ | uds_dogstatsd_to_api_cpu | % cpu utilization | +0.66 | [-0.07, +1.40] | 1 | Logs |
| ➖ | tcp_syslog_to_blackhole | ingress throughput | +0.57 | [+0.52, +0.63] | 1 | Logs |
| ➖ | file_to_blackhole_1000ms_latency_linear_load | egress throughput | +0.24 | [-0.22, +0.70] | 1 | Logs |
| ➖ | file_tree | memory utilization | +0.23 | [+0.10, +0.35] | 1 | Logs |
| ➖ | file_to_blackhole_100ms_latency | egress throughput | +0.10 | [-0.60, +0.81] | 1 | Logs |
| ➖ | file_to_blackhole_500ms_latency | egress throughput | +0.05 | [-0.71, +0.81] | 1 | Logs |
| ➖ | file_to_blackhole_0ms_latency_http2 | egress throughput | +0.01 | [-0.84, +0.86] | 1 | Logs |
| ➖ | tcp_dd_logs_filter_exclude | ingress throughput | -0.00 | [-0.01, +0.01] | 1 | Logs |
| ➖ | uds_dogstatsd_to_api | ingress throughput | -0.00 | [-0.10, +0.10] | 1 | Logs |
| ➖ | file_to_blackhole_300ms_latency | egress throughput | -0.03 | [-0.66, +0.60] | 1 | Logs |
| ➖ | file_to_blackhole_0ms_latency | egress throughput | -0.04 | [-0.88, +0.81] | 1 | Logs |
| ➖ | file_to_blackhole_0ms_latency_http1 | egress throughput | -0.05 | [-0.89, +0.79] | 1 | Logs |
| ➖ | quality_gate_idle | memory utilization | -0.33 | [-0.37, -0.29] | 1 | Logs bounds checks dashboard |
| ➖ | file_to_blackhole_1000ms_latency | egress throughput | -0.42 | [-1.20, +0.36] | 1 | Logs |
| ➖ | quality_gate_logs | % cpu utilization | -0.56 | [-3.49, +2.36] | 1 | Logs |
Bounds Checks: ❌ Failed
| perf | experiment | bounds_check_name | replicates_passed | links |
|---|---|---|---|---|
| ❌ | file_to_blackhole_0ms_latency_http2 | lost_bytes | 7/10 | |
| ❌ | file_to_blackhole_100ms_latency | lost_bytes | 9/10 | |
| ✅ | file_to_blackhole_0ms_latency | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency_http1 | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency_http1 | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency_http2 | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_1000ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_1000ms_latency_linear_load | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_100ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_300ms_latency | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_300ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_500ms_latency | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_500ms_latency | memory_usage | 10/10 | |
| ✅ | quality_gate_idle | memory_usage | 10/10 | bounds checks dashboard |
| ✅ | quality_gate_idle_all_features | memory_usage | 10/10 | bounds checks dashboard |
| ✅ | quality_gate_logs | lost_bytes | 10/10 | |
| ✅ | quality_gate_logs | memory_usage | 10/10 |
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".
CI Pass/Fail Decision
✅ Passed. All Quality Gates passed.
- quality_gate_idle, bounds check memory_usage: 10/10 replicas passed. Gate passed.
- quality_gate_idle_all_features, bounds check memory_usage: 10/10 replicas passed. Gate passed.
- quality_gate_logs, bounds check memory_usage: 10/10 replicas passed. Gate passed.
- quality_gate_logs, bounds check lost_bytes: 10/10 replicas passed. Gate passed.
nhulston
force-pushed
the
nicholas.hulston/s3-span-pointers
branch
from
5d9bc7b to
f804396
Compare
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 |
Uncompressed package size comparisonComparison with ancestor Diff per package
Decision✅ Passed |
What does this PR do?
Adds span pointers to spans for Lambdas triggered by S3
putObject,copyObject, andcompleteMultipartUploadevents. This change will affect the downstream case for Universal Instrumentation Lambda runtimes that still use the legacy Go agent (Java, .NET, Golang).Span pointers are similar to Span Links, but for cases when it is impossible to pass the Trace ID and Span ID between the spans that need to be linked.
When the calculated hashes for the upstream and downstream lambdas match, the Datadog backend will automatically link the two traces together.
Motivation
Span pointers is a new feature being developed by the Serverless team. This feature already exists in Python & Node, and I'm working on adding it to universal instrumentation runtimes (.NET, Java, Golang).
Describe how you validated your changes
Mostly manual testing, but I also added unit tests.
To verify my changes, there are two ways.
Easy: Checkout this span, which is an upstream Lambda that makes all three S3 requests and triggers 3 separate downstream Lambda invocations. Therefore, the 3 span pointers should link to 3 different traces. Each downstream trace should point back to this span.
More thorough: Run this Lambda function (only on the serverless AWS account) with the event payload
{ "shouldPutObject": false, "shouldCopyObject": false, "shouldMultipartUpload": false, }and change one of the bools to true. Check Datadog to ensure that the spans are properly linked.
I also added unit tests:
invoke test --targets=./pkg/serverless/invocationlifecyclePossible Drawbacks / Trade-offs
Additional Notes