[apm] standardize peer tag aggregation

[jdgumz]

What does this PR do?

Builds on previous work by standardizing on a set of default peer tags over which to aggregate.
The tags themselves are an approved list of fields that are then converted to peer.* tags on the backend for use with trace metrics.

The previous peer_tags configuration can still be used to provide supplementary tags if necessary, but we intend for this flag to only be used in exceptional cases. This field cannot be used to supply arbitrary tags, as all tags ultimately are vetted in the backend.

Motivation

To make it easier for customers to adopt peer tags.

Additional Notes

Possible Drawbacks / Trade-offs

Describe how to test/QA your changes

Please ask me for the testing doc.

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.

[maycmlee]

Left a couple of suggestions

[jdgumz]

Left a couple of suggestions

Thank you May! I have committed your suggestions.

[pr-commenter]

Bloop Bleep... Dogbot Here

Regression Detector Results

Run ID: 6decd9ed-9188-4a29-bd96-d3a4746cd3dc
Baseline: 1bf8059
Comparison: 0bcc362
Total datadog-agent CPUs: 7

Explanation

A regression test is an integrated performance test for datadog-agent in a repeatable rig, with varying configuration for datadog-agent. What follows is a statistical summary of a brief datadog-agent run for each configuration across SHAs given above. The goal of these tests are to determine quickly if datadog-agent performance is changed and to what degree by a pull request.

Because a target's optimization goal performance in each experiment will vary somewhat each time it is run, we can only estimate mean differences in optimization goal relative to the baseline target. We express these differences as a percentage change relative to the baseline target, denoted "Δ mean %". These estimates are made to a precision that balances accuracy and cost control. We represent this precision as a 90.00% confidence interval denoted "Δ mean % CI": there is a 90.00% chance that the true value of "Δ mean %" is in that interval.

We decide whether a change in performance is a "regression" -- a change worth investigating further -- if both of the following two criteria are true:

1. The estimated |Δ mean %| ≥ 5.00%. This criterion intends to answer the question "Does the estimated change in mean optimization goal performance have a meaningful impact on your customers?". We assume that when |Δ mean %| < 5.00%, the impact on your customers is not meaningful. We also assume that a performance change in optimization goal is worth investigating whether it is an increase or decrease, so long as the magnitude of the change is sufficiently large.

2. Zero is not in the 90.00% confidence interval "Δ mean % CI" about "Δ mean %". This statement is equivalent to saying that there is at least a 90.00% chance that the mean difference in optimization goal is not zero. This criterion intends to answer the question, "Is there a statistically significant difference in mean optimization goal performance?". It also means there is no more than a 10.00% chance this criterion reports a statistically significant difference when the true difference in mean optimization goal is zero -- a "false positive". We assume you are willing to accept a 10.00% chance of inaccurately detecting a change in performance when no true difference exists.

The table below, if present, lists those experiments that have experienced a statistically significant change in mean optimization goal performance between baseline and comparison SHAs with 90.00% confidence OR have been detected as newly erratic. Negative values of "Δ mean %" mean that baseline is faster, whereas positive values of "Δ mean %" mean that comparison is faster. Results that do not exhibit more than a ±5.00% change in their mean optimization goal are discarded. An experiment is erratic if its coefficient of variation is greater than 0.1. The abbreviated table will be omitted if no interesting change is observed.

No interesting changes in experiment optimization goals with confidence ≥ 90.00% and |Δ mean %| ≥ 5.00%.

Fine details of change detection per experiment.

experiment	goal	Δ mean %	Δ mean % CI	confidence
tcp_syslog_to_blackhole	ingress throughput	+1.92	[+1.78, +2.05]	100.00%
process_agent_real_time_mode	egress throughput	+0.82	[-1.70, +3.35]	40.90%
otel_to_otel_logs	ingress throughput	+0.33	[-1.26, +1.91]	26.58%
file_to_blackhole	egress throughput	+0.16	[-0.28, +0.60]	44.33%
trace_agent_json	ingress throughput	+0.03	[-0.10, +0.16]	32.80%
trace_agent_msgpack	ingress throughput	+0.03	[-0.09, +0.16]	32.98%
file_tree	egress throughput	+0.03	[-1.83, +1.89]	2.04%
dogstatsd_string_interner_8MiB_100k	ingress throughput	+0.02	[-0.02, +0.06]	52.05%
uds_dogstatsd_to_api	ingress throughput	+0.01	[-0.16, +0.19]	10.88%
dogstatsd_string_interner_8MiB_100	ingress throughput	+0.01	[-0.12, +0.14]	10.93%
tcp_dd_logs_filter_exclude	ingress throughput	+0.01	[-0.05, +0.06]	16.52%
dogstatsd_string_interner_64MiB_1k	ingress throughput	+0.00	[-0.13, +0.13]	2.09%
dogstatsd_string_interner_128MiB_1k	ingress throughput	+0.00	[-0.14, +0.14]	0.68%
dogstatsd_string_interner_8MiB_50k	ingress throughput	+0.00	[-0.04, +0.04]	0.00%
dogstatsd_string_interner_64MiB_100	ingress throughput	-0.00	[-0.14, +0.14]	0.18%
dogstatsd_string_interner_128MiB_100	ingress throughput	-0.00	[-0.14, +0.14]	0.31%
dogstatsd_string_interner_8MiB_1k	ingress throughput	-0.00	[-0.10, +0.10]	0.69%
idle	egress throughput	-0.01	[-2.47, +2.46]	0.33%
dogstatsd_string_interner_8MiB_10k	ingress throughput	-0.02	[-0.08, +0.04]	44.75%
process_agent_standard_check	egress throughput	-0.20	[-3.73, +3.32]	7.48%
process_agent_standard_check_with_stats	egress throughput	-0.42	[-2.43, +1.59]	26.73%

[jeremy-hanna]

✅ for agent-shared-component owned files

[knusbaum]

Presumably we recommend enabling compute_stats_by_span_kind if peer_tags_aggregation is enabled?

Do we recommend enabling peer_tags_aggregation?

Will people reading this know what it means?

[jdgumz]

I'll rephrase. The idea here is that if you're using peer_tags_aggregation, you will likely also want the span kind flag enabled too.

[knusbaum]

What should they be looking for in their instrumentation? Can we link them any documentation here?

[jdgumz]

I can provide some more guidance in this comment. We do not have clear documentation that would speak to this specific concern.