datadog-agent icon indicating copy to clipboard operation
datadog-agent copied to clipboard

Published 20 hours ago • verified publisher icon DataDog

[CONTINT-1093] Run kind E2E for multiple k8s versions in parallel

Open adel121 opened this issue 1 year ago • 6 comments

What does this PR do?

Allows running kind E2E tests for multiple kubernetes versions in parallel.

Motivation

Speedup E2E testing for multiple versions of kubernetes.

There are 3 ways to run the kindvm e2e tests:

  • Default behaviour:
    • if no kubernetes version is specified, the tests will be run in parallel for the following versions "1.27", "1.23", "1.20", "1.19".
    • Example: inv new-e2e-tests.run --targets ./tests/containers/ --run TestKindSuite
  • Matrix Version:
    • You can run the invoke command with KUBERNETES_VERSIONS env var set to a comma separated string of versions you wish to run the tests for in parallel.
    • Example: KUBERNETES_VERSIONS=1.29,1.25 inv new-e2e-tests.run --targets ./tests/containers/ --run TestKindSuite
  • Single Version:
    • You can run the E2E tests for a single kubernetes version by setting the -c ddinfra:kubernetesVersion config param in the cli
    • Example: inv new-e2e-tests.run --targets ./tests/containers/ --run TestKindSuite -c ddinfra:kubernetesVersion=1.29

Additional Notes

Possible Drawbacks / Trade-offs

  • Using the current approach, the test logs emitted by pulumi will be a bit messy. However, at the end of the test, we will have a full report of all test results for each tested kubernetes version. For example, running the test for 1.27, 1.26 and 1.25 produced the following report:
--- FAIL: TestKindSuite (0.00s)
    --- FAIL: TestKindSuite/TestKind_1.27 (967.22s)
        --- PASS: TestKindSuite/TestKind_1.27/Test00UpAndRunning (11.21s)
            --- PASS: TestKindSuite/TestKind_1.27/Test00UpAndRunning/agent_pods_are_ready_and_not_restarting (11.21s)
        --- PASS: TestKindSuite/TestKind_1.27/TestCPU (48.93s)
            --- PASS: TestKindSuite/TestKind_1.27/TestCPU/metric___container.cpu.usage{kube_deployment:stress-ng,kube_namespace:workload-cpustress} (12.94s)
            --- PASS: TestKindSuite/TestKind_1.27/TestCPU/metric___container.cpu.limit{kube_deployment:stress-ng,kube_namespace:workload-cpustress} (11.92s)
            --- PASS: TestKindSuite/TestKind_1.27/TestCPU/metric___kubernetes.cpu.usage.total{kube_deployment:stress-ng,kube_namespace:workload-cpustress} (11.45s)
            --- PASS: TestKindSuite/TestKind_1.27/TestCPU/metric___kubernetes.cpu.limits{kube_deployment:stress-ng,kube_namespace:workload-cpustress} (12.61s)
        --- PASS: TestKindSuite/TestKind_1.27/TestDogstatsdInAgent (25.24s)
            --- PASS: TestKindSuite/TestKind_1.27/TestDogstatsdInAgent/metric___custom.metric{kube_deployment:dogstatsd-uds,kube_namespace:workload-dogstatsd} (13.27s)
            --- PASS: TestKindSuite/TestKind_1.27/TestDogstatsdInAgent/metric___custom.metric{kube_deployment:dogstatsd-udp,kube_namespace:workload-dogstatsd} (11.96s)
        --- PASS: TestKindSuite/TestKind_1.27/TestDogstatsdStandalone (25.66s)
            --- PASS: TestKindSuite/TestKind_1.27/TestDogstatsdStandalone/metric___custom.metric{kube_deployment:dogstatsd-uds,kube_namespace:workload-dogstatsd-standalone} (11.83s)
            --- PASS: TestKindSuite/TestKind_1.27/TestDogstatsdStandalone/metric___custom.metric{kube_deployment:dogstatsd-udp,kube_namespace:workload-dogstatsd-standalone} (13.83s)
        --- PASS: TestKindSuite/TestKind_1.27/TestNginx (369.78s)
            --- PASS: TestKindSuite/TestKind_1.27/TestNginx/metric___nginx.net.request_per_s{} (21.79s)
            --- PASS: TestKindSuite/TestKind_1.27/TestNginx/metric___network.http.response_time{} (18.19s)
            --- PASS: TestKindSuite/TestKind_1.27/TestNginx/metric___kubernetes_state.deployment.replicas_available{kube_deployment:nginx,kube_namespace:workload-nginx} (15.21s)
            --- PASS: TestKindSuite/TestKind_1.27/TestNginx/log___apps-nginx-server{} (10.91s)
            --- PASS: TestKindSuite/TestKind_1.27/TestNginx/hpa___kubernetes_state.deployment.replicas_available{kube_namespace:workload-nginx,kube_deployment:nginx} (303.67s)
        --- PASS: TestKindSuite/TestKind_1.27/TestPrometheus (22.21s)
            --- PASS: TestKindSuite/TestKind_1.27/TestPrometheus/metric___prom_gauge{kube_deployment:prometheus,kube_namespace:workload-prometheus} (22.21s)
        --- PASS: TestKindSuite/TestKind_1.27/TestRedis (78.33s)
            --- PASS: TestKindSuite/TestKind_1.27/TestRedis/metric___redis.net.instantaneous_ops_per_sec{} (19.93s)
            --- PASS: TestKindSuite/TestKind_1.27/TestRedis/metric___kubernetes_state.deployment.replicas_available{kube_deployment:redis,kube_namespace:workload-redis} (19.62s)
            --- PASS: TestKindSuite/TestKind_1.27/TestRedis/log___redis{} (11.93s)
            --- PASS: TestKindSuite/TestKind_1.27/TestRedis/hpa___kubernetes_state.deployment.replicas_available{kube_namespace:workload-redis,kube_deployment:redis} (26.85s)
        --- FAIL: TestKindSuite/TestKind_1.27/TestVersion (5.42s)
            --- FAIL: TestKindSuite/TestKind_1.27/TestVersion/Linux_agent_pods_are_running_the_good_version (2.99s)
            --- PASS: TestKindSuite/TestKind_1.27/TestVersion/Windows_agent_pods_are_running_the_good_version (0.13s)
            --- FAIL: TestKindSuite/TestKind_1.27/TestVersion/cluster_agent_pods_are_running_the_good_version (1.15s)
            --- FAIL: TestKindSuite/TestKind_1.27/TestVersion/cluster_checks_pods_are_running_the_good_version (1.15s)
        --- PASS: TestKindSuite/TestKind_1.27/TestZZUpAndRunning (12.72s)
            --- PASS: TestKindSuite/TestKind_1.27/TestZZUpAndRunning/agent_pods_are_ready_and_not_restarting (12.72s)
    --- FAIL: TestKindSuite/TestKind_1.25 (986.71s)
        --- PASS: TestKindSuite/TestKind_1.25/Test00UpAndRunning (11.17s)
            --- PASS: TestKindSuite/TestKind_1.25/Test00UpAndRunning/agent_pods_are_ready_and_not_restarting (11.17s)
        --- PASS: TestKindSuite/TestKind_1.25/TestCPU (56.81s)
            --- PASS: TestKindSuite/TestKind_1.25/TestCPU/metric___container.cpu.usage{kube_deployment:stress-ng,kube_namespace:workload-cpustress} (12.02s)
            --- PASS: TestKindSuite/TestKind_1.25/TestCPU/metric___container.cpu.limit{kube_deployment:stress-ng,kube_namespace:workload-cpustress} (11.50s)
            --- PASS: TestKindSuite/TestKind_1.25/TestCPU/metric___kubernetes.cpu.usage.total{kube_deployment:stress-ng,kube_namespace:workload-cpustress} (21.30s)
            --- PASS: TestKindSuite/TestKind_1.25/TestCPU/metric___kubernetes.cpu.limits{kube_deployment:stress-ng,kube_namespace:workload-cpustress} (11.99s)
        --- PASS: TestKindSuite/TestKind_1.25/TestDogstatsdInAgent (25.49s)
            --- PASS: TestKindSuite/TestKind_1.25/TestDogstatsdInAgent/metric___custom.metric{kube_deployment:dogstatsd-uds,kube_namespace:workload-dogstatsd} (13.48s)
            --- PASS: TestKindSuite/TestKind_1.25/TestDogstatsdInAgent/metric___custom.metric{kube_deployment:dogstatsd-udp,kube_namespace:workload-dogstatsd} (12.00s)
        --- PASS: TestKindSuite/TestKind_1.25/TestDogstatsdStandalone (29.15s)
            --- PASS: TestKindSuite/TestKind_1.25/TestDogstatsdStandalone/metric___custom.metric{kube_deployment:dogstatsd-uds,kube_namespace:workload-dogstatsd-standalone} (14.07s)
            --- PASS: TestKindSuite/TestKind_1.25/TestDogstatsdStandalone/metric___custom.metric{kube_deployment:dogstatsd-udp,kube_namespace:workload-dogstatsd-standalone} (15.08s)
        --- PASS: TestKindSuite/TestKind_1.25/TestNginx (380.05s)
            --- PASS: TestKindSuite/TestKind_1.25/TestNginx/metric___nginx.net.request_per_s{} (17.56s)
            --- PASS: TestKindSuite/TestKind_1.25/TestNginx/metric___network.http.response_time{} (17.53s)
            --- PASS: TestKindSuite/TestKind_1.25/TestNginx/metric___kubernetes_state.deployment.replicas_available{kube_deployment:nginx,kube_namespace:workload-nginx} (15.68s)
            --- PASS: TestKindSuite/TestKind_1.25/TestNginx/log___apps-nginx-server{} (13.39s)
            --- PASS: TestKindSuite/TestKind_1.25/TestNginx/hpa___kubernetes_state.deployment.replicas_available{kube_namespace:workload-nginx,kube_deployment:nginx} (315.89s)
        --- PASS: TestKindSuite/TestKind_1.25/TestPrometheus (45.62s)
            --- PASS: TestKindSuite/TestKind_1.25/TestPrometheus/metric___prom_gauge{kube_deployment:prometheus,kube_namespace:workload-prometheus} (45.62s)
        --- PASS: TestKindSuite/TestKind_1.25/TestRedis (76.03s)
            --- PASS: TestKindSuite/TestKind_1.25/TestRedis/metric___redis.net.instantaneous_ops_per_sec{} (18.40s)
            --- PASS: TestKindSuite/TestKind_1.25/TestRedis/metric___kubernetes_state.deployment.replicas_available{kube_deployment:redis,kube_namespace:workload-redis} (27.54s)
            --- PASS: TestKindSuite/TestKind_1.25/TestRedis/log___redis{} (11.62s)
            --- PASS: TestKindSuite/TestKind_1.25/TestRedis/hpa___kubernetes_state.deployment.replicas_available{kube_namespace:workload-redis,kube_deployment:redis} (18.46s)
        --- FAIL: TestKindSuite/TestKind_1.25/TestVersion (4.37s)
            --- FAIL: TestKindSuite/TestKind_1.25/TestVersion/Linux_agent_pods_are_running_the_good_version (1.60s)
            --- PASS: TestKindSuite/TestKind_1.25/TestVersion/Windows_agent_pods_are_running_the_good_version (0.14s)
            --- FAIL: TestKindSuite/TestKind_1.25/TestVersion/cluster_agent_pods_are_running_the_good_version (1.25s)
            --- FAIL: TestKindSuite/TestKind_1.25/TestVersion/cluster_checks_pods_are_running_the_good_version (1.38s)
        --- PASS: TestKindSuite/TestKind_1.25/TestZZUpAndRunning (11.68s)
            --- PASS: TestKindSuite/TestKind_1.25/TestZZUpAndRunning/agent_pods_are_ready_and_not_restarting (11.68s)
    --- FAIL: TestKindSuite/TestKind_1.26 (1031.60s)
        --- PASS: TestKindSuite/TestKind_1.26/Test00UpAndRunning (12.30s)
            --- PASS: TestKindSuite/TestKind_1.26/Test00UpAndRunning/agent_pods_are_ready_and_not_restarting (12.30s)
        --- PASS: TestKindSuite/TestKind_1.26/TestCPU (111.58s)
            --- PASS: TestKindSuite/TestKind_1.26/TestCPU/metric___container.cpu.usage{kube_deployment:stress-ng,kube_namespace:workload-cpustress} (16.35s)
            --- PASS: TestKindSuite/TestKind_1.26/TestCPU/metric___container.cpu.limit{kube_deployment:stress-ng,kube_namespace:workload-cpustress} (14.77s)
            --- PASS: TestKindSuite/TestKind_1.26/TestCPU/metric___kubernetes.cpu.usage.total{kube_deployment:stress-ng,kube_namespace:workload-cpustress} (66.00s)
            --- PASS: TestKindSuite/TestKind_1.26/TestCPU/metric___kubernetes.cpu.limits{kube_deployment:stress-ng,kube_namespace:workload-cpustress} (14.46s)
        --- PASS: TestKindSuite/TestKind_1.26/TestDogstatsdInAgent (27.30s)
            --- PASS: TestKindSuite/TestKind_1.26/TestDogstatsdInAgent/metric___custom.metric{kube_deployment:dogstatsd-uds,kube_namespace:workload-dogstatsd} (13.09s)
            --- PASS: TestKindSuite/TestKind_1.26/TestDogstatsdInAgent/metric___custom.metric{kube_deployment:dogstatsd-udp,kube_namespace:workload-dogstatsd} (14.21s)
        --- FAIL: TestKindSuite/TestKind_1.26/TestDogstatsdStandalone (240.97s)
            --- FAIL: TestKindSuite/TestKind_1.26/TestDogstatsdStandalone/metric___custom.metric{kube_deployment:dogstatsd-uds,kube_namespace:workload-dogstatsd-standalone} (120.46s)
            --- FAIL: TestKindSuite/TestKind_1.26/TestDogstatsdStandalone/metric___custom.metric{kube_deployment:dogstatsd-udp,kube_namespace:workload-dogstatsd-standalone} (120.51s)
        --- PASS: TestKindSuite/TestKind_1.26/TestNginx (106.29s)
            --- PASS: TestKindSuite/TestKind_1.26/TestNginx/metric___nginx.net.request_per_s{} (29.26s)
            --- PASS: TestKindSuite/TestKind_1.26/TestNginx/metric___network.http.response_time{} (24.50s)
            --- PASS: TestKindSuite/TestKind_1.26/TestNginx/metric___kubernetes_state.deployment.replicas_available{kube_deployment:nginx,kube_namespace:workload-nginx} (23.89s)
            --- PASS: TestKindSuite/TestKind_1.26/TestNginx/log___apps-nginx-server{} (11.60s)
            --- PASS: TestKindSuite/TestKind_1.26/TestNginx/hpa___kubernetes_state.deployment.replicas_available{kube_namespace:workload-nginx,kube_deployment:nginx} (17.05s)
        --- PASS: TestKindSuite/TestKind_1.26/TestPrometheus (24.88s)
            --- PASS: TestKindSuite/TestKind_1.26/TestPrometheus/metric___prom_gauge{kube_deployment:prometheus,kube_namespace:workload-prometheus} (24.88s)
        --- PASS: TestKindSuite/TestKind_1.26/TestRedis (68.42s)
            --- PASS: TestKindSuite/TestKind_1.26/TestRedis/metric___redis.net.instantaneous_ops_per_sec{} (18.43s)
            --- PASS: TestKindSuite/TestKind_1.26/TestRedis/metric___kubernetes_state.deployment.replicas_available{kube_deployment:redis,kube_namespace:workload-redis} (19.70s)
            --- PASS: TestKindSuite/TestKind_1.26/TestRedis/log___redis{} (11.60s)
            --- PASS: TestKindSuite/TestKind_1.26/TestRedis/hpa___kubernetes_state.deployment.replicas_available{kube_namespace:workload-redis,kube_deployment:redis} (18.69s)
        --- FAIL: TestKindSuite/TestKind_1.26/TestVersion (4.95s)
            --- FAIL: TestKindSuite/TestKind_1.26/TestVersion/Linux_agent_pods_are_running_the_good_version (2.22s)
            --- PASS: TestKindSuite/TestKind_1.26/TestVersion/Windows_agent_pods_are_running_the_good_version (0.13s)
            --- FAIL: TestKindSuite/TestKind_1.26/TestVersion/cluster_agent_pods_are_running_the_good_version (1.20s)
            --- FAIL: TestKindSuite/TestKind_1.26/TestVersion/cluster_checks_pods_are_running_the_good_version (1.40s)
        --- PASS: TestKindSuite/TestKind_1.26/TestZZUpAndRunning (11.07s)
            --- PASS: TestKindSuite/TestKind_1.26/TestZZUpAndRunning/agent_pods_are_ready_and_not_restarting (11.07s)
  • Potential improvement: currently, we create 1 stack per kubernetes version. It would make sense, from resource optimisation standpoint, to have all resources created within one stack and run the tests for each version. However, this will require handling resources isolation based on kubernetes versions.
  • Potential improvement: the tests for different kubernetes versions may be able to share the same fakeintake. However, currently it is not possible to distinguish sources via the fakeintake.

Describe how to test/QA your changes

Reviewer's Checklist

  • [ ] If known, an appropriate milestone has been selected; otherwise the Triage milestone is set.
  • [ ] Use the major_change label 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.
  • [ ] A release note has been added or the changelog/no-changelog label has been applied.
  • [ ] Changed code has automated tests for its functionality.
  • [ ] Adequate QA/testing plan information is provided if the qa/skip-qa label is not applied.
  • [ ] At least one team/.. label has been applied, indicating the team(s) that should QA this change.
  • [ ] If applicable, docs team has been notified or an issue has been opened on the documentation repo.
  • [ ] If applicable, the need-change/operator and need-change/helm labels have been applied.
  • [ ] If applicable, the k8s/<min-version> label, indicating the lowest Kubernetes version compatible with this feature.
  • [ ] If applicable, the config template has been updated.

adel121 avatar Dec 13 '23 13:12 adel121

TestKindSuite are broken.

L3n41c avatar Dec 13 '23 18:12 L3n41c

TestKindSuite are broken.

@L3n41c

This is weird. The logs include fatal error: concurrent map writes.

When I run the tests with inv new-e2e-tests.run --targets ./tests/containers/ --run TestKindSuite it works fine, and here is a snippet from the output I get:

=== RUN   TestKindSuite/TestKind_1.20/Test00UpAndRunning
=== RUN   TestKindSuite/TestKind_1.20/Test00UpAndRunning/agent_pods_are_ready_and_not_restarting
=== RUN   TestKindSuite/TestKind_1.23/TestCPU
=== RUN   TestKindSuite/TestKind_1.23/TestCPU/metric___container.cpu.usage{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.19/TestCPU
=== RUN   TestKindSuite/TestKind_1.19/TestCPU/metric___container.cpu.usage{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.27/TestCPU/metric___container.cpu.limit{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.20/TestCPU
=== RUN   TestKindSuite/TestKind_1.20/TestCPU/metric___container.cpu.usage{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.23/TestCPU/metric___container.cpu.limit{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.19/TestCPU/metric___container.cpu.limit{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.27/TestCPU/metric___kubernetes.cpu.usage.total{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.23/TestCPU/metric___kubernetes.cpu.usage.total{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.20/TestCPU/metric___container.cpu.limit{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.19/TestCPU/metric___kubernetes.cpu.usage.total{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.27/TestCPU/metric___kubernetes.cpu.limits{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.20/TestCPU/metric___kubernetes.cpu.usage.total{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.27/TestControlPlane
=== RUN   TestKindSuite/TestKind_1.27/TestControlPlane/metric___kube_apiserver.apiserver_request_total{}
=== RUN   TestKindSuite/TestKind_1.27/TestControlPlane/metric___kube_controller_manager.queue.adds{}
=== RUN   TestKindSuite/TestKind_1.27/TestControlPlane/metric___kube_scheduler.schedule_attempts{}
=== RUN   TestKindSuite/TestKind_1.20/TestCPU/metric___kubernetes.cpu.limits{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.23/TestCPU/metric___kubernetes.cpu.limits{kube_deployment:stress-ng,kube_namespace:workload-cpustress}
=== RUN   TestKindSuite/TestKind_1.27/TestDogstatsdInAgent
=== RUN   TestKindSuite/TestKind_1.27/TestDogstatsdInAgent/metric___custom.metric{kube_deployment:dogstatsd-uds,kube_namespace:workload-dogstatsd}
=== RUN   TestKindSuite/TestKind_1.20/TestControlPlane

Need to check why it says fatal error: concurrent map writes when run within a gitlab job.

adel121 avatar Dec 13 '23 19:12 adel121

TestKindSuite are broken.

I restarted the job, and it used the same stacks that had been created in the previous run.

It didn't break, and the output is as expected.

Some assertions failed on some specific kubernetes versions, but this is unrelated to the PR.

TBH I have no idea why the first run of the job emitted fatal error: concurrent map writes

adel121 avatar Dec 13 '23 20:12 adel121

The fatal error: concurrent map writes errors were coming from the fact that the fake intake client isn’t thread-safe. This will eventually be fixed by #21507.

But the failures in the TestControlPlane tests are due to the kind version change.

L3n41c avatar Dec 14 '23 08:12 L3n41c

Moved milestone to 7.52 given that it's just test optimization

AliDatadog avatar Jan 05 '24 11:01 AliDatadog

Bloop Bleep... Dogbot Here

Regression Detector Results

Run ID: 8672fb10-e0ad-454d-986a-0480c9357cc1 Baseline: cbb39b7eeb80f5377e54a574a23eb13e56f8db77 Comparison: 2d3c4e9bdcc59ff56e4b14c7164324a338337877

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

Experiments with missing or malformed data

  • basic_py_check

Usually, this warning means that there is no usable optimization goal data for that experiment, which could be a result of misconfiguration.

No significant changes in experiment optimization goals

Confidence level: 90.00% Effect size tolerance: |Δ mean %| ≥ 5.00%

There were no significant changes in experiment optimization goals at this confidence level and effect size tolerance.

Experiments ignored for regressions

Regressions in experiments with settings containing erratic: true are ignored.

perf experiment goal Δ mean % Δ mean % CI
➖ file_to_blackhole % cpu utilization +0.64 [-5.93, +7.21]

Fine details of change detection per experiment

perf experiment goal Δ mean % Δ mean % CI
➖ uds_dogstatsd_to_api_cpu % cpu utilization +3.35 [+1.91, +4.79]
➖ process_agent_standard_check memory utilization +0.72 [+0.69, +0.76]
➖ file_to_blackhole % cpu utilization +0.64 [-5.93, +7.21]
➖ otel_to_otel_logs ingress throughput +0.62 [-0.02, +1.26]
➖ idle memory utilization +0.24 [+0.21, +0.27]
➖ file_tree memory utilization +0.20 [+0.12, +0.27]
➖ process_agent_real_time_mode memory utilization +0.03 [-0.00, +0.06]
➖ uds_dogstatsd_to_api ingress throughput +0.00 [-0.00, +0.00]
➖ tcp_dd_logs_filter_exclude ingress throughput -0.00 [-0.00, +0.00]
➖ trace_agent_json ingress throughput -0.00 [-0.04, +0.04]
➖ trace_agent_msgpack ingress throughput -0.01 [-0.02, +0.01]
➖ tcp_syslog_to_blackhole ingress throughput -0.02 [-0.08, +0.03]
➖ process_agent_standard_check_with_stats memory utilization -0.22 [-0.25, -0.19]

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:

  1. Its estimated |Δ mean %| ≥ 5.00%, indicating the change is big enough to merit a closer look.

  2. 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.

  3. Its configuration does not mark it "erratic".

pr-commenter[bot] avatar Jan 22 '24 18:01 pr-commenter[bot]

Closing in favour of this PR

adel121 avatar Mar 18 '24 08:03 adel121