fix(loki): serialize the structured metadata to JSON

[maxboone]

fixes: #21443

Structured metadata wasn't added to the serializer implementation, which likely caused it not to be sent to Loki. I added the serialization.

Considering the Encoder<Vec<LokiRecord>> for LokiBatchEncoder is not unit tested anywhere, thus I tested it by trying to (JSON) serialize the event and see if the structured metadata is (correctly) in there.

[jszwedko]

Nice! Thank you @maxboone !

[github-actions]

Regression Detector Results

Run ID: dadd0a54-2ce2-4b87-ac9f-64af883392b1 Metrics dashboard

Baseline: e250dcb
Comparison: 4ba4d61

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_regex_logs2metric_ddmetrics	ingress throughput	-5.54	[-5.74, -5.34]
❌	socket_to_socket_blackhole	ingress throughput	-7.48	[-7.55, -7.40]
❌	otlp_http_to_blackhole	ingress throughput	-9.97	[-10.12, -9.83]

Experiments ignored for regressions

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

perf	experiment	goal	Δ mean %	Δ mean % CI	links
➖	file_to_blackhole	egress throughput	-4.28	[-10.99, +2.44]

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	links
➖	http_elasticsearch	ingress throughput	+1.92	[+1.75, +2.09]
➖	datadog_agent_remap_blackhole	ingress throughput	+0.83	[+0.66, +1.01]
➖	splunk_hec_route_s3	ingress throughput	+0.82	[+0.52, +1.13]
➖	datadog_agent_remap_datadog_logs	ingress throughput	+0.56	[+0.34, +0.77]
➖	http_to_s3	ingress throughput	+0.51	[+0.24, +0.79]
➖	fluent_elasticsearch	ingress throughput	+0.35	[-0.16, +0.86]
➖	http_to_http_acks	ingress throughput	+0.32	[-0.91, +1.55]
➖	http_to_http_noack	ingress throughput	+0.16	[+0.08, +0.25]
➖	http_to_http_json	ingress throughput	+0.02	[-0.02, +0.05]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	+0.01	[-0.08, +0.11]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	+0.01	[-0.10, +0.11]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	+0.00	[-0.08, +0.09]
➖	http_text_to_http_json	ingress throughput	-0.58	[-0.71, -0.45]
➖	syslog_loki	ingress throughput	-0.68	[-0.74, -0.61]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	-0.80	[-1.03, -0.57]
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	-0.89	[-1.01, -0.77]
➖	otlp_grpc_to_blackhole	ingress throughput	-1.39	[-1.50, -1.28]
➖	syslog_log2metric_humio_metrics	ingress throughput	-1.40	[-1.59, -1.22]
➖	datadog_agent_remap_blackhole_acks	ingress throughput	-2.29	[-2.51, -2.06]
➖	syslog_log2metric_splunk_hec_metrics	ingress throughput	-2.43	[-2.57, -2.28]
➖	syslog_splunk_hec_logs	ingress throughput	-2.43	[-2.58, -2.29]
➖	syslog_humio_logs	ingress throughput	-3.02	[-3.15, -2.88]
➖	file_to_blackhole	egress throughput	-4.28	[-10.99, +2.44]
❌	syslog_regex_logs2metric_ddmetrics	ingress throughput	-5.54	[-5.74, -5.34]
❌	socket_to_socket_blackhole	ingress throughput	-7.48	[-7.55, -7.40]
❌	otlp_http_to_blackhole	ingress throughput	-9.97	[-10.12, -9.83]

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]

Regression Detector Results

Run ID: f3cb2a9e-0c84-46e5-a780-4c5da82af27f Metrics dashboard

Baseline: b3dac6e
Comparison: c6d45a1

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
✅	datadog_agent_remap_blackhole_acks	ingress throughput	+6.31	[+6.16, +6.47]
❌	socket_to_socket_blackhole	ingress throughput	-10.53	[-10.59, -10.47]
❌	otlp_http_to_blackhole	ingress throughput	-11.91	[-12.06, -11.76]

Experiments ignored for regressions

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

perf	experiment	goal	Δ mean %	Δ mean % CI	links
➖	file_to_blackhole	egress throughput	-3.79	[-10.70, +3.11]

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	links
✅	datadog_agent_remap_blackhole_acks	ingress throughput	+6.31	[+6.16, +6.47]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	+2.67	[+2.52, +2.83]
➖	datadog_agent_remap_datadog_logs	ingress throughput	+2.48	[+2.28, +2.69]
➖	syslog_regex_logs2metric_ddmetrics	ingress throughput	+1.53	[+1.38, +1.68]
➖	fluent_elasticsearch	ingress throughput	+1.22	[+0.72, +1.72]
➖	syslog_log2metric_humio_metrics	ingress throughput	+1.05	[+0.91, +1.19]
➖	http_to_http_acks	ingress throughput	+0.59	[-0.64, +1.83]
➖	http_to_s3	ingress throughput	+0.49	[+0.22, +0.76]
➖	splunk_hec_route_s3	ingress throughput	+0.43	[+0.10, +0.77]
➖	http_to_http_noack	ingress throughput	+0.12	[+0.05, +0.19]
➖	http_to_http_json	ingress throughput	+0.04	[-0.01, +0.09]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	+0.03	[-0.07, +0.12]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	+0.02	[-0.07, +0.11]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	-0.01	[-0.09, +0.08]
➖	syslog_log2metric_splunk_hec_metrics	ingress throughput	-1.90	[-2.03, -1.77]
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	-2.01	[-2.12, -1.91]
➖	otlp_grpc_to_blackhole	ingress throughput	-2.46	[-2.57, -2.34]
➖	syslog_loki	ingress throughput	-2.53	[-2.62, -2.44]
➖	http_text_to_http_json	ingress throughput	-2.53	[-2.72, -2.33]
➖	datadog_agent_remap_blackhole	ingress throughput	-2.88	[-3.03, -2.73]
➖	syslog_splunk_hec_logs	ingress throughput	-2.93	[-3.06, -2.80]
➖	http_elasticsearch	ingress throughput	-3.13	[-3.31, -2.94]
➖	syslog_humio_logs	ingress throughput	-3.28	[-3.44, -3.13]
➖	file_to_blackhole	egress throughput	-3.79	[-10.70, +3.11]
❌	socket_to_socket_blackhole	ingress throughput	-10.53	[-10.59, -10.47]
❌	otlp_http_to_blackhole	ingress throughput	-11.91	[-12.06, -11.76]

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]

Regression Detector Results

Run ID: 92cdced0-c2bb-4178-a799-50064aafeb14 Metrics dashboard

Baseline: 202da8b
Comparison: 0eb9efc

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_regex_logs2metric_ddmetrics	ingress throughput	+7.92	[+7.69, +8.15]

Experiments ignored for regressions

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

perf	experiment	goal	Δ mean %	Δ mean % CI	links
➖	file_to_blackhole	egress throughput	+6.85	[-0.33, +14.03]

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	links
✅	syslog_regex_logs2metric_ddmetrics	ingress throughput	+7.92	[+7.69, +8.15]
➖	file_to_blackhole	egress throughput	+6.85	[-0.33, +14.03]
➖	syslog_log2metric_splunk_hec_metrics	ingress throughput	+4.67	[+4.48, +4.85]
➖	http_text_to_http_json	ingress throughput	+2.37	[+2.06, +2.68]
➖	datadog_agent_remap_datadog_logs	ingress throughput	+1.62	[+1.42, +1.83]
➖	syslog_log2metric_tag_cardinality_limit_blackhole	ingress throughput	+0.82	[+0.70, +0.93]
➖	splunk_hec_route_s3	ingress throughput	+0.72	[+0.42, +1.03]
➖	syslog_log2metric_humio_metrics	ingress throughput	+0.72	[+0.49, +0.95]
➖	http_to_s3	ingress throughput	+0.40	[+0.13, +0.67]
➖	datadog_agent_remap_blackhole_acks	ingress throughput	+0.37	[+0.27, +0.47]
➖	http_to_http_noack	ingress throughput	+0.15	[+0.07, +0.23]
➖	http_to_http_json	ingress throughput	+0.07	[+0.01, +0.14]
➖	splunk_hec_to_splunk_hec_logs_noack	ingress throughput	+0.01	[-0.09, +0.10]
➖	splunk_hec_indexer_ack_blackhole	ingress throughput	-0.00	[-0.08, +0.08]
➖	syslog_humio_logs	ingress throughput	-0.00	[-0.14, +0.13]
➖	splunk_hec_to_splunk_hec_logs_acks	ingress throughput	-0.03	[-0.14, +0.09]
➖	syslog_loki	ingress throughput	-0.04	[-0.15, +0.06]
➖	http_elasticsearch	ingress throughput	-0.11	[-0.30, +0.08]
➖	syslog_splunk_hec_logs	ingress throughput	-0.35	[-0.43, -0.27]
➖	socket_to_socket_blackhole	ingress throughput	-0.59	[-0.64, -0.53]
➖	otlp_grpc_to_blackhole	ingress throughput	-0.64	[-0.77, -0.52]
➖	fluent_elasticsearch	ingress throughput	-0.78	[-1.26, -0.30]
➖	datadog_agent_remap_blackhole	ingress throughput	-1.11	[-1.27, -0.96]
➖	http_to_http_acks	ingress throughput	-1.63	[-2.84, -0.42]
➖	datadog_agent_remap_datadog_logs_acks	ingress throughput	-1.69	[-1.87, -1.51]
➖	otlp_http_to_blackhole	ingress throughput	-4.55	[-4.69, -4.41]

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".