chore(smp upgrade) - Upgrade SMP to latest release by scottopell · Pull Request #20713 · vectordotdev/vector

scottopell · 2024-06-21T21:33:47Z

No description provided.

bits-bot · 2024-06-21T21:33:51Z

All committers have signed the CLA.

scottopell · 2024-06-21T21:33:58Z

/ci-run-regression

datadog-vectordotdev · 2024-06-21T21:58:39Z

Datadog Report

Branch report: smp/official-0-15-1-upgrade
Commit report: 1fbaa50
Test service: vector

✅ 0 Failed, 7 Passed, 0 Skipped, 25.42s Total Time

…een lading and target

scottopell · 2024-06-21T22:03:13Z

/ci-run-regression

scottopell · 2024-06-24T14:30:07Z

/ci-run-regression

.github/workflows/regression.yml

github-actions · 2024-06-24T18:49:34Z

Regression Detector Results

Run ID: 5ca95b9b-f8b0-42ae-a6f7-9a46d8fa7e61 Metrics dashboard

Baseline: 52759aa
Comparison: 61a8824

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

Significant changes in experiment optimization goals

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

perf	experiment	goal	Δ mean %	Δ mean % CI	links
❌	file_to_blackhole	egress throughput	-10.19	[-16.73, -3.65]

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI
➖	syslog_humio_logs	ingress throughput	+1.64	[+1.49, +1.80]
➖	syslog_splunk_hec_logs	ingress throughput	+0.52	[+0.40, +0.64]
➖	http_elasticsearch	ingress throughput	+0.15	[-0.01, +0.31]
➖	splunk_hec_route_s3	ingress throughput	+0.14	[-0.23, +0.51]
➖	http_to_http_noack	ingress throughput	+0.08	[+0.02, +0.14]
➖	http_to_http_json	ingress throughput	+0.04	[-0.00, +0.09]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	+0.01	[-0.07, +0.09]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	-0.00	[-0.10, +0.09]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	-0.00	[-0.13, +0.12]
➖	otlp_http_to_blackhole	ingress throughput	-0.06	[-0.22, +0.09]
➖	datadog_agent_remap_datadog_logs	ingress throughput	-0.10	[-0.31, +0.12]
➖	datadog_agent_remap_blackhole_acks	ingress throughput	-0.18	[-0.31, -0.06]
➖	otlp_grpc_to_blackhole	ingress throughput	-0.23	[-0.35, -0.11]
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	-0.23	[-0.34, -0.12]
➖	socket_to_socket_blackhole	ingress throughput	-0.32	[-0.38, -0.25]
➖	http_to_s3	ingress throughput	-0.37	[-0.65, -0.10]
➖	syslog_log2metric_splunk_hec_metrics	ingress throughput	-0.93	[-1.02, -0.84]
➖	syslog_regex_logs2metric_ddmetrics	ingress throughput	-0.94	[-1.12, -0.75]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	-1.49	[-1.67, -1.30]
➖	datadog_agent_remap_blackhole	ingress throughput	-1.50	[-1.63, -1.37]
➖	fluent_elasticsearch	ingress throughput	-1.52	[-2.00, -1.04]
➖	http_text_to_http_json	ingress throughput	-2.12	[-2.25, -1.99]
➖	syslog_loki	ingress throughput	-2.13	[-2.23, -2.04]
➖	http_to_http_acks	ingress throughput	-2.33	[-3.63, -1.03]
➖	syslog_log2metric_humio_metrics	ingress throughput	-3.51	[-3.64, -3.37]
❌	file_to_blackhole	egress throughput	-10.19	[-16.73, -3.65]

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".

.github/workflows/workload_checks.yml

GeorgeHahn · 2024-06-24T20:55:51Z

regression/cases/fluent_elasticsearch/lading/lading.yaml

      seed: [2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47, 53,
        59, 61, 67, 71, 73, 79, 83, 89, 97, 101, 103, 107, 109, 113, 127, 131]
-      addr: "localhost:8282"
+      addr: "0.0.0.0:8282"


I'm somewhat surprised that 0.0.0.0 works here; lading is connecting to this and I didn't know Linux supported connections to that address. (I'd have expected 127.0.0.1 here.) No change needed, just curious.

It is possible that this specific one wasn't needed, I just took a "big hammer" approach and made everything 0.0.0.0

I'm also a bit surprised by this. Did you find this change necessary? I could see it causing confusion in the future. I'd also expect localhost here.

I'll try with localhost in a followup PR to see.

The hostname localhost is broken inside target containers (never investigated; it's a longstanding bug that'd be really nice to have fixed.)

GeorgeHahn · 2024-06-24T20:56:24Z

regression/config.yaml

+# Additional variables for per-experiment links:
+# - `experiment`: the name of the experiment
+report:
+  metrics_dashboard: "https://app.datadoghq.com/dashboard/ykh-ua8-vcu/SMP-Regression-Detector-Metrics?fromUser=true&refresh_mode=paused&tpl_var_run-id%5B0%5D={{ job_id }}&view=spans&from_ts={{ start_time_ms }}&to_ts={{ end_time_ms }}&live=false"


Would you mind also adding a logs link?

jszwedko

I left a few nits/questions, but overall 👍

jszwedko · 2024-06-24T21:00:17Z

.github/workflows/regression.yml

-      lading-version: ${{ steps.experimental-meta.outputs.LADING_VERSION }}
    steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4


We actually withheld this version bump in #18490 but I think it's ok to do for this workflow.

jszwedko · 2024-06-24T21:00:59Z

regression/cases/datadog_agent_remap_blackhole/experiment.yaml

+
+

Suggested change

Nit: extra whitespace 😄

jszwedko · 2024-06-24T21:02:35Z

regression/cases/fluent_elasticsearch/lading/lading.yaml

      seed: [2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47, 53,
        59, 61, 67, 71, 73, 79, 83, 89, 97, 101, 103, 107, 109, 113, 127, 131]
-      addr: "localhost:8282"
+      addr: "0.0.0.0:8282"


I'm also a bit surprised by this. Did you find this change necessary? I could see it causing confusion in the future. I'd also expect localhost here.

github-actions · 2024-06-24T21:52:12Z

Regression Detector Results

Run ID: d19fc108-249e-44bd-940e-5d7232856a2a Metrics dashboard

Baseline: 52759aa
Comparison: 5e7981c

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

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.

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI
➖	syslog_log2metric_splunk_hec_metrics	ingress throughput	+1.17	[+1.06, +1.28]
➖	fluent_elasticsearch	ingress throughput	+1.13	[+0.63, +1.62]
➖	splunk_hec_route_s3	ingress throughput	+1.01	[+0.61, +1.42]
➖	http_text_to_http_json	ingress throughput	+0.95	[+0.83, +1.07]
➖	syslog_regex_logs2metric_ddmetrics	ingress throughput	+0.54	[+0.38, +0.71]
➖	http_to_http_json	ingress throughput	+0.11	[+0.05, +0.18]
➖	http_to_http_noack	ingress throughput	+0.09	[+0.03, +0.16]
➖	syslog_loki	ingress throughput	+0.06	[-0.02, +0.13]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	+0.00	[-0.10, +0.10]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	-0.00	[-0.08, +0.08]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	-0.00	[-0.13, +0.12]
➖	http_to_s3	ingress throughput	-0.01	[-0.28, +0.26]
➖	otlp_grpc_to_blackhole	ingress throughput	-0.10	[-0.21, +0.01]
➖	otlp_http_to_blackhole	ingress throughput	-0.13	[-0.25, +0.00]
➖	http_to_http_acks	ingress throughput	-0.13	[-1.44, +1.17]
➖	datadog_agent_remap_blackhole_acks	ingress throughput	-0.34	[-0.49, -0.18]
➖	datadog_agent_remap_datadog_logs	ingress throughput	-0.38	[-0.61, -0.15]
➖	socket_to_socket_blackhole	ingress throughput	-0.85	[-0.89, -0.80]
➖	syslog_splunk_hec_logs	ingress throughput	-0.88	[-1.04, -0.73]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	-1.00	[-1.20, -0.81]
➖	datadog_agent_remap_blackhole	ingress throughput	-1.08	[-1.21, -0.96]
➖	syslog_log2metric_humio_metrics	ingress throughput	-1.18	[-1.34, -1.03]
➖	syslog_humio_logs	ingress throughput	-1.79	[-1.89, -1.68]
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	-1.91	[-2.03, -1.79]
➖	http_elasticsearch	ingress throughput	-2.75	[-2.94, -2.56]
➖	file_to_blackhole	egress throughput	-5.12	[-12.09, +1.85]

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".

* Update smp job submission and associated versions * Adds experiment.yamls and config.yaml * Remove VECTOR_REQUIRE_HEALTHY to better accomodate startup races between lading and target * format * Removes unused dotfile-wannabe * Attempt to run off PR * Eval should_run always * Get PR number correctly for PRs * New merge-base cmd * New git ref for merge-base cmd * Only checkout PR if its not a PR * Full ref to master * Show me all valid git refs * Fetch full git repo * Skip merge-equeue metadata step for pull request * normalize to 0.0.0.0 instead of localhost * Normalize to 0.0.0.0 * Revert changes to manually run this on PR * Removes workload checks * Removes trailing empty whitespace lines --------- Co-authored-by: George Hahn <george.hahn@datadoghq.com>

scottopell added 2 commits June 21, 2024 17:32

Update smp job submission and associated versions

1f052d3

Adds experiment.yamls and config.yaml

9d55f17

scottopell mentioned this pull request Jun 21, 2024

chore(upgrade smp): Update to SMP 0.15.1 #20712

Closed

scottopell added the no-changelog Changes in this PR do not need user-facing explanations in the release changelog label Jun 21, 2024

Remove VECTOR_REQUIRE_HEALTHY to better accomodate startup races betw…

bc2cc28

…een lading and target

GeorgeHahn and others added 2 commits June 21, 2024 16:16

format

be7022e

Removes unused dotfile-wannabe

67837c5

scottopell added 6 commits June 24, 2024 11:05

Attempt to run off PR

d0b9109

Eval should_run always

e7fa40b

Get PR number correctly for PRs

f9cf41b

New merge-base cmd

86b31b2

New git ref for merge-base cmd

f6caaab

Only checkout PR if its not a PR

90419e9

github-advanced-security bot found potential problems Jun 24, 2024

View reviewed changes

.github/workflows/regression.yml Fixed Show fixed Hide fixed

scottopell added 6 commits June 24, 2024 11:27

Full ref to master

50f6cba

Show me all valid git refs

7baa079

Fetch full git repo

300d363

Skip merge-equeue metadata step for pull request

5d0e192

normalize to 0.0.0.0 instead of localhost

5203845

Normalize to 0.0.0.0

366a9ae

Revert changes to manually run this on PR

c2d6570

GeorgeHahn reviewed Jun 24, 2024

View reviewed changes

.github/workflows/workload_checks.yml Show resolved Hide resolved

Removes workload checks

ccf585f

GeorgeHahn reviewed Jun 24, 2024

View reviewed changes

GeorgeHahn approved these changes Jun 24, 2024

View reviewed changes

jszwedko approved these changes Jun 24, 2024

View reviewed changes

Removes trailing empty whitespace lines

97350cf

scottopell enabled auto-merge June 24, 2024 21:09

scottopell added this pull request to the merge queue Jun 24, 2024

Merged via the queue into master with commit 5e7981c Jun 24, 2024

scottopell deleted the smp/official-0-15-1-upgrade branch June 24, 2024 22:38

Conversation

scottopell commented Jun 21, 2024

Uh oh!

bits-bot commented Jun 21, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

scottopell commented Jun 21, 2024

Uh oh!

datadog-vectordotdev bot commented Jun 21, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Datadog Report

Uh oh!

scottopell commented Jun 21, 2024

Uh oh!

scottopell commented Jun 24, 2024

Uh oh!

Uh oh!

github-actions bot commented Jun 24, 2024

Regression Detector Results

Significant changes in experiment optimization goals

Fine details of change detection per experiment

Explanation

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jszwedko left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Jun 24, 2024

Regression Detector Results

No significant changes in experiment optimization goals

Fine details of change detection per experiment

Explanation

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

bits-bot commented Jun 21, 2024 •

edited

Loading

datadog-vectordotdev bot commented Jun 21, 2024 •

edited

Loading