[NDMII-2360] Replace the profile map with a profile provider

[dplepage-dd]

What does this PR do?

This PR changes the SNMP check to use a profile.Provider type for looking up profiles instead of a plain map[string]ProfileConfig. The overall behavior of the integration shouldn't change, as the only implementation of Provider in this PR is a static one that just provides profiles from a map[string]ProfileConfig.

Motivation

This is part of converting the SNMP integration to receive profiles via Remote Configuration - the first step is ensuring that profiles are fetched from a controlled source instead of being redundantly cached in a variety of places, so that I can add a new Provider type that listens for RC updates and rebuilds the internal configuration mapping.

Describe how to test/QA your changes

Unit tests have been updated; the behavior of the agent shouldn't change.

[cit-pr-commenter]

Regression Detector

Regression Detector Results

Metrics dashboard
Target profiles
Run ID: 9454df83-e95a-497d-93f7-cfcc1ed2a6fb

Baseline: 681517f
Comparison: 6ad5b9d
Diff

Optimization Goals: ✅ No significant changes detected

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
➖	basic_py_check	% cpu utilization	+2.68	[-1.28, +6.64]	1	Logs
➖	quality_gate_idle_all_features	memory utilization	+1.42	[+1.29, +1.54]	1	Logs bounds checks dashboard
➖	file_tree	memory utilization	+1.22	[+1.07, +1.36]	1	Logs
➖	file_to_blackhole_1000ms_latency	egress throughput	+0.43	[-0.37, +1.23]	1	Logs
➖	otel_to_otel_logs	ingress throughput	+0.30	[-0.40, +1.00]	1	Logs
➖	file_to_blackhole_500ms_latency	egress throughput	+0.09	[-0.67, +0.85]	1	Logs
➖	file_to_blackhole_1000ms_latency_linear_load	egress throughput	+0.03	[-0.43, +0.50]	1	Logs
➖	file_to_blackhole_300ms_latency	egress throughput	+0.02	[-0.61, +0.66]	1	Logs
➖	uds_dogstatsd_to_api	ingress throughput	+0.00	[-0.10, +0.10]	1	Logs
➖	tcp_dd_logs_filter_exclude	ingress throughput	-0.00	[-0.01, +0.01]	1	Logs
➖	file_to_blackhole_100ms_latency	egress throughput	-0.02	[-0.75, +0.71]	1	Logs
➖	file_to_blackhole_0ms_latency	egress throughput	-0.10	[-1.01, +0.81]	1	Logs
➖	quality_gate_idle	memory utilization	-0.49	[-0.54, -0.45]	1	Logs bounds checks dashboard
➖	pycheck_lots_of_tags	% cpu utilization	-0.70	[-4.08, +2.69]	1	Logs
➖	uds_dogstatsd_to_api_cpu	% cpu utilization	-0.74	[-1.47, -0.00]	1	Logs
➖	tcp_syslog_to_blackhole	ingress throughput	-0.93	[-1.00, -0.86]	1	Logs

Bounds Checks: ❌ Failed

perf	experiment	bounds_check_name	replicates_passed	links
❌	file_to_blackhole_0ms_latency	lost_bytes	8/10
❌	file_to_blackhole_500ms_latency	lost_bytes	9/10
✅	file_to_blackhole_0ms_latency	memory_usage	10/10
✅	file_to_blackhole_1000ms_latency	memory_usage	10/10
✅	file_to_blackhole_1000ms_latency_linear_load	memory_usage	10/10
✅	file_to_blackhole_100ms_latency	lost_bytes	10/10
✅	file_to_blackhole_100ms_latency	memory_usage	10/10
✅	file_to_blackhole_300ms_latency	lost_bytes	10/10
✅	file_to_blackhole_300ms_latency	memory_usage	10/10
✅	file_to_blackhole_500ms_latency	memory_usage	10/10
✅	quality_gate_idle	memory_usage	10/10	bounds checks dashboard
✅	quality_gate_idle_all_features	memory_usage	10/10	bounds checks dashboard

Explanation

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

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

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".

CI Pass/Fail Decision

✅ Passed. All Quality Gates passed.

• quality_gate_idle, bounds check memory_usage: 10/10 replicas passed. Gate passed.
• quality_gate_idle_all_features, bounds check memory_usage: 10/10 replicas passed. Gate passed.

[agent-platform-auto-pr]

Test changes on VM

Use this command from test-infra-definitions to manually test this PR changes on a VM:

inv aws.create-vm --pipeline-id=50868614 --os-family=ubuntu

Note: This applies to commit e1c9a66

[FlorianVeaux]

The PR looks great to me, nice job on it !
I've just left a few questions and naming questions

[FlorianVeaux]

Should we have a custom GetProfileForSysObjectID instead ?
This avoid the repetition of getting the profile name, then getting the profile with a potential error which should never happen

[dplepage-dd]

The behavior right now is:

1. If a profile name is hard-coded in the config, we set .ProfileName when the config is loaded
2. When we start looking for metrics, if there isn't a hard-coded profile, we fetch the sysobjectid and then set .ProfileName to the name of the profile that best matches the sysobjectid.
3. In both cases, after setting .ProfileName, we fetch the profile in question from the provider and recompute config.OidConfig based on it.

So if we want to fetch the profile directly instead of having the name as an intermediate step, we'll still need to handle the case where it's been hardcoded in the config.

Now that I think about it, though, it would be a good idea to do that - we shouldn't be caching the name of the best matching profile, because the best match could change when RC updates.

[dplepage-dd]

Making this change proved a little complicated, so I'm going to put it in a separate PR (I'm removing the caching fields .Metrics, .MetricTags, .ProfileTags, .Metadata, and .OidConfig from the CheckConfig type and making them a separate object that is held by the DeviceCheck and recomputed when it detects that the sysobjectid or the profile manager has changed).

[FlorianVeaux]

❓ question: ‏There's a slight risk here (and in GetProfile) that the caller can modify the content of the profiles. Did you see it and would you think it's worth to mitigate it?

[dplepage-dd]

For GetAllProfiles it seemed excessive to clone them all; now that we don't have that anymore, it's just GetProfile that has this risk. I didn't want to copy them because I thought it would be wasteful to copy a profile every time we access it, but maybe that's just me thinking too much like a Python programmer - probably the time needed to duplicate a profile is tiny enough that we can just go ahead and do it every time. Do you think it would be better to just duplicate the profile every time GetProfile() is called?

[agent-platform-auto-pr]

Package size comparison

Comparison with ancestor 3c882e71c8b1830d40be8d4cfae580e40414d207

Diff per package

package	diff	status	size	ancestor	threshold
datadog-agent-amd64-deb	-0.01MB	✅	1265.85MB	1265.86MB	140.00MB
datadog-iot-agent-amd64-deb	-0.02MB	✅	113.26MB	113.28MB	10.00MB
datadog-dogstatsd-amd64-deb	0.00MB	✅	78.41MB	78.41MB	10.00MB
datadog-heroku-agent-amd64-deb	-0.03MB	✅	526.62MB	526.65MB	70.00MB
datadog-agent-x86_64-rpm	-0.01MB	✅	1275.09MB	1275.10MB	140.00MB
datadog-agent-x86_64-suse	-0.01MB	✅	1275.09MB	1275.10MB	140.00MB
datadog-iot-agent-x86_64-rpm	-0.02MB	✅	113.33MB	113.35MB	10.00MB
datadog-iot-agent-x86_64-suse	-0.02MB	✅	113.33MB	113.35MB	10.00MB
datadog-dogstatsd-x86_64-rpm	0.00MB	✅	78.49MB	78.49MB	10.00MB
datadog-dogstatsd-x86_64-suse	0.00MB	✅	78.49MB	78.49MB	10.00MB
datadog-agent-arm64-deb	-0.02MB	✅	1000.85MB	1000.87MB	140.00MB
datadog-iot-agent-arm64-deb	-0.02MB	✅	108.74MB	108.77MB	10.00MB
datadog-dogstatsd-arm64-deb	0.00MB	✅	55.64MB	55.64MB	10.00MB
datadog-agent-aarch64-rpm	-0.02MB	✅	1010.06MB	1010.08MB	140.00MB
datadog-iot-agent-aarch64-rpm	-0.02MB	✅	108.81MB	108.84MB	10.00MB

Decision

✅ Passed

[dplepage-dd]

/merge

[dd-devflow]

Devflow running: /merge

View all feedbacks in Devflow UI.

---

2024-12-11 22:08:42 UTC ℹ️ MergeQueue: pull request added to the queue

The median merge time in main is 26m.

---

2024-12-11 22:34:53 UTC ℹ️ MergeQueue: This merge request was merged