enhancement(datadog_agent source): enable zstd decompression

[StephenWakely]

trafficstars

The datadog agent is in the process of allowing payloads to be compressed with zstd. This adds the ability to Vector to handle these.

[datadog-vectordotdev[bot]]

Datadog Report

Branch report: stephen/agent-metrics-zstd Commit report: 9c2735d Test service: vector

:white_check_mark: 0 Failed, 11 Passed, 0 Skipped, 25.45s Total Time

[github-actions[bot]]

Regression Detector Results

Run ID: 4beaa9f7-f52c-4cbf-a047-a2cb299c0e07 Metrics dashboard

Baseline: 3de6f0b8175c8a08da1974d6b6e9634075244a06 Comparison: 4da65ce1d1dc55518f48ee6a416673c84b648df9

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	-23.94	[-29.89, -17.99]

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	links
➖	http_elasticsearch	ingress throughput	+2.39	[+2.21, +2.57]
➖	otlp_http_to_blackhole	ingress throughput	+1.94	[+1.82, +2.07]
➖	datadog_agent_remap_blackhole	ingress throughput	+1.77	[+1.65, +1.89]
➖	http_text_to_http_json	ingress throughput	+1.23	[+1.11, +1.34]
➖	otlp_grpc_to_blackhole	ingress throughput	+1.13	[+1.01, +1.25]
➖	syslog_splunk_hec_logs	ingress throughput	+0.45	[+0.34, +0.57]
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	+0.37	[+0.26, +0.48]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	+0.14	[-0.06, +0.34]
➖	http_to_http_json	ingress throughput	+0.10	[+0.04, +0.17]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	+0.00	[-0.08, +0.09]
➖	http_to_http_noack	ingress throughput	-0.00	[-0.02, +0.02]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	-0.00	[-0.10, +0.09]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	-0.00	[-0.10, +0.10]
➖	splunk_hec_route_s3	ingress throughput	-0.03	[-0.39, +0.33]
➖	http_to_s3	ingress throughput	-0.08	[-0.35, +0.19]
➖	syslog_loki	ingress throughput	-0.41	[-0.49, -0.32]
➖	http_to_http_acks	ingress throughput	-0.44	[-1.76, +0.88]
➖	datadog_agent_remap_datadog_logs	ingress throughput	-0.74	[-0.95, -0.53]
➖	socket_to_socket_blackhole	ingress throughput	-0.74	[-0.80, -0.69]
➖	fluent_elasticsearch	ingress throughput	-1.11	[-1.61, -0.62]
➖	datadog_agent_remap_blackhole_acks	ingress throughput	-1.26	[-1.36, -1.16]
➖	syslog_humio_logs	ingress throughput	-1.37	[-1.50, -1.24]
➖	syslog_regex_logs2metric_ddmetrics	ingress throughput	-1.47	[-1.67, -1.26]
➖	syslog_log2metric_humio_metrics	ingress throughput	-1.98	[-2.10, -1.86]
➖	syslog_log2metric_splunk_hec_metrics	ingress throughput	-2.74	[-2.85, -2.64]
❌	file_to_blackhole	egress throughput	-23.94	[-29.89, -17.99]

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".

[github-actions[bot]]

Regression Detector Results

Run ID: 87402219-2f5c-48c4-bf90-90de69cf803b Metrics dashboard

Baseline: 3de6f0b8175c8a08da1974d6b6e9634075244a06 Comparison: 328495597e9c7388f887cc4e0725b8e24e7da4ce

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	links
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	+3.56	[+3.44, +3.67]
➖	otlp_http_to_blackhole	ingress throughput	+2.53	[+2.39, +2.66]
➖	datadog_agent_remap_blackhole	ingress throughput	+2.26	[+2.13, +2.39]
➖	http_elasticsearch	ingress throughput	+1.64	[+1.48, +1.80]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	+1.40	[+1.16, +1.64]
➖	http_text_to_http_json	ingress throughput	+1.39	[+1.24, +1.54]
➖	http_to_s3	ingress throughput	+0.66	[+0.38, +0.94]
➖	syslog_log2metric_humio_metrics	ingress throughput	+0.65	[+0.53, +0.78]
➖	splunk_hec_route_s3	ingress throughput	+0.45	[+0.05, +0.84]
➖	fluent_elasticsearch	ingress throughput	+0.43	[-0.07, +0.92]
➖	otlp_grpc_to_blackhole	ingress throughput	+0.36	[+0.24, +0.47]
➖	http_to_http_json	ingress throughput	+0.12	[+0.05, +0.19]
➖	http_to_http_noack	ingress throughput	+0.06	[+0.01, +0.11]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	+0.01	[-0.09, +0.10]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	+0.01	[-0.09, +0.11]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	-0.02	[-0.10, +0.05]
➖	file_to_blackhole	egress throughput	-0.20	[-6.97, +6.57]
➖	syslog_log2metric_splunk_hec_metrics	ingress throughput	-0.67	[-0.82, -0.52]
➖	syslog_splunk_hec_logs	ingress throughput	-0.83	[-0.94, -0.71]
➖	http_to_http_acks	ingress throughput	-0.84	[-2.15, +0.47]
➖	datadog_agent_remap_datadog_logs	ingress throughput	-1.10	[-1.29, -0.91]
➖	socket_to_socket_blackhole	ingress throughput	-1.16	[-1.23, -1.09]
➖	syslog_regex_logs2metric_ddmetrics	ingress throughput	-1.52	[-1.69, -1.35]
➖	datadog_agent_remap_blackhole_acks	ingress throughput	-1.84	[-1.96, -1.72]
➖	syslog_humio_logs	ingress throughput	-2.11	[-2.21, -2.01]
➖	syslog_loki	ingress throughput	-2.19	[-2.27, -2.11]

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".

[github-actions[bot]]

Regression Detector Results

Run ID: 61139ac6-7da1-41ae-bcde-e478ea7a29f9 Metrics dashboard

Baseline: 3de6f0b8175c8a08da1974d6b6e9634075244a06 Comparison: c62b975c84e2da779976e233ed893c5461980735

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
❌	syslog_log2metric_splunk_hec_metrics	ingress throughput	-6.94	[-7.09, -6.80]

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	links
➖	http_elasticsearch	ingress throughput	+2.66	[+2.49, +2.83]
➖	otlp_http_to_blackhole	ingress throughput	+2.34	[+2.21, +2.47]
➖	otlp_grpc_to_blackhole	ingress throughput	+1.45	[+1.33, +1.57]
➖	syslog_humio_logs	ingress throughput	+1.23	[+1.07, +1.38]
➖	http_to_s3	ingress throughput	+0.56	[+0.29, +0.83]
➖	syslog_splunk_hec_logs	ingress throughput	+0.49	[+0.37, +0.62]
➖	socket_to_socket_blackhole	ingress throughput	+0.35	[+0.30, +0.41]
➖	splunk_hec_route_s3	ingress throughput	+0.16	[-0.21, +0.53]
➖	http_to_http_noack	ingress throughput	+0.06	[+0.01, +0.12]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	+0.04	[-0.03, +0.12]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	+0.04	[-0.14, +0.22]
➖	http_to_http_json	ingress throughput	+0.03	[-0.01, +0.07]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	+0.00	[-0.10, +0.11]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	-0.00	[-0.10, +0.10]
➖	http_text_to_http_json	ingress throughput	-0.00	[-0.14, +0.14]
➖	datadog_agent_remap_datadog_logs	ingress throughput	-0.08	[-0.31, +0.15]
➖	syslog_log2metric_humio_metrics	ingress throughput	-0.37	[-0.49, -0.24]
➖	fluent_elasticsearch	ingress throughput	-0.70	[-1.19, -0.21]
➖	file_to_blackhole	egress throughput	-0.82	[-7.46, +5.82]
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	-0.98	[-1.08, -0.87]
➖	datadog_agent_remap_blackhole_acks	ingress throughput	-1.02	[-1.15, -0.89]
➖	syslog_loki	ingress throughput	-1.16	[-1.25, -1.07]
➖	http_to_http_acks	ingress throughput	-1.79	[-3.09, -0.49]
➖	datadog_agent_remap_blackhole	ingress throughput	-2.03	[-2.15, -1.90]
➖	syslog_regex_logs2metric_ddmetrics	ingress throughput	-3.75	[-3.90, -3.60]
❌	syslog_log2metric_splunk_hec_metrics	ingress throughput	-6.94	[-7.09, -6.80]

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".

[github-actions[bot]]

Regression Detector Results

Run ID: 8132d56b-68fa-4704-b0fc-d7562f2d526b Metrics dashboard

Baseline: 3de6f0b8175c8a08da1974d6b6e9634075244a06 Comparison: 687fa966b3ed09b38c4e7aef5c3794954e5c37ad

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
❌	syslog_humio_logs	ingress throughput	-5.11	[-5.29, -4.93]

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	links
➖	file_to_blackhole	egress throughput	+4.46	[-2.41, +11.33]
➖	splunk_hec_route_s3	ingress throughput	+3.62	[+3.20, +4.03]
➖	otlp_http_to_blackhole	ingress throughput	+1.64	[+1.47, +1.81]
➖	http_elasticsearch	ingress throughput	+0.93	[+0.76, +1.11]
➖	syslog_splunk_hec_logs	ingress throughput	+0.87	[+0.76, +0.97]
➖	http_to_http_acks	ingress throughput	+0.38	[-0.93, +1.68]
➖	datadog_agent_remap_datadog_logs	ingress throughput	+0.37	[+0.16, +0.57]
➖	syslog_log2metric_humio_metrics	ingress throughput	+0.19	[+0.06, +0.33]
➖	http_to_http_json	ingress throughput	+0.08	[+0.02, +0.14]
➖	http_to_http_noack	ingress throughput	+0.06	[+0.01, +0.11]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	+0.02	[-0.07, +0.12]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	+0.01	[-0.10, +0.11]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	-0.00	[-0.10, +0.10]
➖	datadog_agent_remap_blackhole_acks	ingress throughput	-0.05	[-0.15, +0.06]
➖	otlp_grpc_to_blackhole	ingress throughput	-0.05	[-0.16, +0.06]
➖	fluent_elasticsearch	ingress throughput	-0.09	[-0.59, +0.42]
➖	datadog_agent_remap_blackhole	ingress throughput	-0.19	[-0.32, -0.07]
➖	http_text_to_http_json	ingress throughput	-0.27	[-0.42, -0.13]
➖	socket_to_socket_blackhole	ingress throughput	-0.30	[-0.37, -0.23]
➖	syslog_regex_logs2metric_ddmetrics	ingress throughput	-0.39	[-0.57, -0.22]
➖	http_to_s3	ingress throughput	-0.51	[-0.78, -0.24]
➖	syslog_log2metric_splunk_hec_metrics	ingress throughput	-0.75	[-0.87, -0.64]
➖	syslog_loki	ingress throughput	-0.96	[-1.05, -0.87]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	-1.04	[-1.19, -0.89]
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	-1.17	[-1.31, -1.04]
❌	syslog_humio_logs	ingress throughput	-5.11	[-5.29, -4.93]

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".

[github-actions[bot]]

Regression Detector Results

Run ID: e13874f0-68e3-485b-be0e-c74ca0e85c71 Metrics dashboard

Baseline: 951d726a70df7a2045524e7ef9517fdc6423d9f2 Comparison: 245a91e8e78730919d1dfa34892181e851ff9f84

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	-21.88	[-28.08, -15.67]

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	links
➖	otlp_http_to_blackhole	ingress throughput	+1.94	[+1.81, +2.08]
➖	syslog_log2metric_splunk_hec_metrics	ingress throughput	+1.77	[+1.64, +1.90]
➖	syslog_log2metric_humio_metrics	ingress throughput	+1.73	[+1.59, +1.87]
➖	datadog_agent_remap_datadog_logs	ingress throughput	+1.07	[+0.84, +1.30]
➖	http_elasticsearch	ingress throughput	+0.98	[+0.77, +1.20]
➖	otlp_grpc_to_blackhole	ingress throughput	+0.87	[+0.76, +0.99]
➖	datadog_agent_remap_blackhole_acks	ingress throughput	+0.83	[+0.72, +0.95]
➖	http_to_s3	ingress throughput	+0.48	[+0.20, +0.75]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	+0.42	[+0.31, +0.54]
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	+0.41	[+0.32, +0.51]
➖	fluent_elasticsearch	ingress throughput	+0.04	[-0.45, +0.54]
➖	http_to_http_json	ingress throughput	+0.04	[-0.00, +0.09]
➖	http_to_http_noack	ingress throughput	+0.02	[-0.01, +0.05]
➖	splunk_hec_route_s3	ingress throughput	+0.00	[-0.37, +0.38]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	+0.00	[-0.08, +0.08]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	-0.00	[-0.10, +0.10]
➖	http_to_http_acks	ingress throughput	-0.07	[-1.38, +1.24]
➖	datadog_agent_remap_blackhole	ingress throughput	-0.09	[-0.20, +0.02]
➖	http_text_to_http_json	ingress throughput	-1.50	[-1.62, -1.38]
➖	socket_to_socket_blackhole	ingress throughput	-1.60	[-1.65, -1.55]
➖	syslog_humio_logs	ingress throughput	-1.71	[-1.91, -1.50]
➖	syslog_loki	ingress throughput	-1.77	[-1.87, -1.67]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	-2.51	[-2.73, -2.28]
➖	syslog_regex_logs2metric_ddmetrics	ingress throughput	-2.96	[-3.13, -2.78]
➖	syslog_splunk_hec_logs	ingress throughput	-3.23	[-3.41, -3.04]
❌	file_to_blackhole	egress throughput	-21.88	[-28.08, -15.67]

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".

[github-actions[bot]]

Regression Detector Results

Run ID: 35bd2809-e5f1-4dbb-985a-0f96500d6e12 Metrics dashboard

Baseline: 951d726a70df7a2045524e7ef9517fdc6423d9f2 Comparison: 73f659afb212fe88b31c6b5ee8c0ba24b147d7cd

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	-15.28	[-21.88, -8.67]

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	links
➖	syslog_log2metric_splunk_hec_metrics	ingress throughput	+2.96	[+2.83, +3.09]
➖	otlp_grpc_to_blackhole	ingress throughput	+1.25	[+1.13, +1.37]
➖	http_elasticsearch	ingress throughput	+1.23	[+1.08, +1.38]
➖	otlp_http_to_blackhole	ingress throughput	+1.01	[+0.87, +1.15]
➖	syslog_regex_logs2metric_ddmetrics	ingress throughput	+0.95	[+0.71, +1.19]
➖	fluent_elasticsearch	ingress throughput	+0.78	[+0.29, +1.27]
➖	datadog_agent_remap_blackhole_acks	ingress throughput	+0.36	[+0.24, +0.47]
➖	http_to_http_noack	ingress throughput	+0.12	[+0.05, +0.19]
➖	http_to_http_json	ingress throughput	+0.10	[+0.04, +0.17]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	+0.00	[-0.09, +0.09]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	-0.00	[-0.12, +0.12]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	-0.00	[-0.08, +0.08]
➖	syslog_log2metric_humio_metrics	ingress throughput	-0.04	[-0.17, +0.09]
➖	datadog_agent_remap_blackhole	ingress throughput	-0.37	[-0.46, -0.27]
➖	http_to_s3	ingress throughput	-0.40	[-0.67, -0.14]
➖	syslog_splunk_hec_logs	ingress throughput	-0.42	[-0.54, -0.31]
➖	http_text_to_http_json	ingress throughput	-0.54	[-0.67, -0.42]
➖	socket_to_socket_blackhole	ingress throughput	-0.66	[-0.73, -0.60]
➖	splunk_hec_route_s3	ingress throughput	-1.21	[-1.59, -0.83]
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	-1.33	[-1.46, -1.21]
➖	http_to_http_acks	ingress throughput	-1.49	[-2.80, -0.18]
➖	syslog_loki	ingress throughput	-1.64	[-1.71, -1.58]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	-1.70	[-1.91, -1.49]
➖	datadog_agent_remap_datadog_logs	ingress throughput	-1.79	[-2.01, -1.56]
➖	syslog_humio_logs	ingress throughput	-3.54	[-3.67, -3.42]
❌	file_to_blackhole	egress throughput	-15.28	[-21.88, -8.67]

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".

[github-actions[bot]]

Regression Detector Results

Run ID: f3f4b83f-8e24-4b04-96d9-fbaf182ccac4 Metrics dashboard

Baseline: 252951154ce2300568539b4e37fc3f03c829c18d Comparison: 8bfeaa194fcebca6b51a89b604de372490e4b3c4

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	+9.19	[+1.79, +16.58]

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	links
✅	file_to_blackhole	egress throughput	+9.19	[+1.79, +16.58]
➖	syslog_regex_logs2metric_ddmetrics	ingress throughput	+3.08	[+2.90, +3.25]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	+2.93	[+2.68, +3.18]
➖	syslog_log2metric_splunk_hec_metrics	ingress throughput	+2.12	[+2.00, +2.24]
➖	syslog_humio_logs	ingress throughput	+1.85	[+1.69, +2.01]
➖	splunk_hec_route_s3	ingress throughput	+1.46	[+1.06, +1.87]
➖	otlp_http_to_blackhole	ingress throughput	+1.16	[+1.01, +1.30]
➖	socket_to_socket_blackhole	ingress throughput	+1.10	[+1.02, +1.17]
➖	syslog_splunk_hec_logs	ingress throughput	+1.03	[+0.93, +1.14]
➖	http_to_http_noack	ingress throughput	+0.07	[+0.02, +0.12]
➖	http_to_http_json	ingress throughput	+0.06	[+0.01, +0.11]
➖	fluent_elasticsearch	ingress throughput	+0.05	[-0.44, +0.55]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	+0.05	[-0.03, +0.12]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	-0.00	[-0.10, +0.10]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	-0.01	[-0.11, +0.10]
➖	syslog_loki	ingress throughput	-0.25	[-0.34, -0.16]
➖	http_to_s3	ingress throughput	-0.37	[-0.64, -0.09]
➖	datadog_agent_remap_blackhole_acks	ingress throughput	-0.48	[-0.60, -0.37]
➖	http_to_http_acks	ingress throughput	-0.79	[-2.10, +0.53]
➖	otlp_grpc_to_blackhole	ingress throughput	-0.92	[-1.04, -0.81]
➖	datadog_agent_remap_blackhole	ingress throughput	-1.06	[-1.18, -0.95]
➖	datadog_agent_remap_datadog_logs	ingress throughput	-1.73	[-1.99, -1.48]
➖	http_elasticsearch	ingress throughput	-1.85	[-2.02, -1.67]
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	-2.64	[-2.73, -2.54]
➖	http_text_to_http_json	ingress throughput	-2.69	[-2.84, -2.54]
➖	syslog_log2metric_humio_metrics	ingress throughput	-3.41	[-3.56, -3.27]

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".

[github-actions[bot]]

Regression Detector Results

Run ID: f548e530-3f50-423c-83d0-13f4fc261982 Metrics dashboard

Baseline: e982f6679e9d1526d856efa462e728774b52cf34 Comparison: b77a871bc0c882c8889265dbcc62b8dfed566b52

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	-8.66	[-15.30, -2.03]

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	links
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	+3.81	[+3.71, +3.91]
➖	syslog_log2metric_splunk_hec_metrics	ingress throughput	+1.72	[+1.63, +1.82]
➖	syslog_loki	ingress throughput	+1.47	[+1.38, +1.56]
➖	socket_to_socket_blackhole	ingress throughput	+0.83	[+0.77, +0.88]
➖	otlp_http_to_blackhole	ingress throughput	+0.72	[+0.58, +0.85]
➖	otlp_grpc_to_blackhole	ingress throughput	+0.42	[+0.30, +0.54]
➖	datadog_agent_remap_blackhole_acks	ingress throughput	+0.19	[+0.03, +0.35]
➖	splunk_hec_route_s3	ingress throughput	+0.11	[-0.29, +0.51]
➖	http_to_http_json	ingress throughput	+0.07	[+0.01, +0.13]
➖	fluent_elasticsearch	ingress throughput	+0.05	[-0.44, +0.54]
➖	http_to_http_noack	ingress throughput	+0.05	[-0.00, +0.10]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	+0.00	[-0.09, +0.09]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	-0.01	[-0.14, +0.12]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	-0.01	[-0.10, +0.07]
➖	datadog_agent_remap_blackhole	ingress throughput	-0.17	[-0.29, -0.06]
➖	http_to_s3	ingress throughput	-0.26	[-0.54, +0.01]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	-0.29	[-0.49, -0.10]
➖	http_text_to_http_json	ingress throughput	-0.36	[-0.50, -0.22]
➖	syslog_splunk_hec_logs	ingress throughput	-0.37	[-0.47, -0.27]
➖	http_elasticsearch	ingress throughput	-0.85	[-1.03, -0.67]
➖	syslog_regex_logs2metric_ddmetrics	ingress throughput	-1.31	[-1.48, -1.14]
➖	syslog_log2metric_humio_metrics	ingress throughput	-1.32	[-1.47, -1.18]
➖	syslog_humio_logs	ingress throughput	-1.55	[-1.64, -1.46]
➖	http_to_http_acks	ingress throughput	-1.72	[-3.02, -0.41]
➖	datadog_agent_remap_datadog_logs	ingress throughput	-2.44	[-2.65, -2.22]
❌	file_to_blackhole	egress throughput	-8.66	[-15.30, -2.03]

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".

[github-actions[bot]]

Regression Detector Results

Run ID: 2a05b23d-3503-4f84-ad75-0a78350f0986 Metrics dashboard

Baseline: d778c6f016f862a2d05ba8fda435d0c9cceb2715 Comparison: def8b7f251ec60acc4ec673f5c6acc4065e99e18

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	links
➖	datadog_agent_remap_blackhole_acks	ingress throughput	+3.25	[+3.13, +3.37]
➖	splunk_hec_route_s3	ingress throughput	+2.32	[+1.93, +2.70]
➖	http_elasticsearch	ingress throughput	+0.90	[+0.72, +1.08]
➖	syslog_log2metric_splunk_hec_metrics	ingress throughput	+0.88	[+0.77, +1.00]
➖	datadog_agent_remap_blackhole	ingress throughput	+0.73	[+0.64, +0.83]
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	+0.10	[-0.00, +0.20]
➖	http_to_http_json	ingress throughput	+0.04	[-0.00, +0.09]
➖	http_to_http_noack	ingress throughput	+0.03	[-0.01, +0.07]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	+0.00	[-0.08, +0.08]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	-0.00	[-0.11, +0.11]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	-0.00	[-0.10, +0.10]
➖	http_to_s3	ingress throughput	-0.15	[-0.42, +0.11]
➖	fluent_elasticsearch	ingress throughput	-0.32	[-0.81, +0.17]
➖	otlp_grpc_to_blackhole	ingress throughput	-0.39	[-0.50, -0.28]
➖	syslog_log2metric_humio_metrics	ingress throughput	-0.68	[-0.81, -0.56]
➖	socket_to_socket_blackhole	ingress throughput	-0.99	[-1.05, -0.93]
➖	file_to_blackhole	egress throughput	-1.07	[-7.61, +5.47]
➖	http_to_http_acks	ingress throughput	-1.33	[-2.63, -0.02]
➖	syslog_regex_logs2metric_ddmetrics	ingress throughput	-1.49	[-1.70, -1.29]
➖	datadog_agent_remap_datadog_logs	ingress throughput	-1.70	[-1.91, -1.49]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	-1.87	[-2.07, -1.67]
➖	syslog_loki	ingress throughput	-2.23	[-2.32, -2.14]
➖	otlp_http_to_blackhole	ingress throughput	-2.29	[-2.43, -2.15]
➖	syslog_splunk_hec_logs	ingress throughput	-2.78	[-2.88, -2.67]
➖	syslog_humio_logs	ingress throughput	-2.99	[-3.10, -2.88]
➖	http_text_to_http_json	ingress throughput	-3.24	[-3.35, -3.13]

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".