fix: merge gauges properly and avoid dropping existing points

[tobz]

Context

When metrics are merged in the aggregate transform, we call MetricValues::merge, which is meant to uniformly handle the merge of different points for all metric types. This method handles both the merging of points based on their timestamp, as well as the specific merge behavior for the point values themselves.

An issue was pointed out with how ADP was handling gauges, and after some debugging, it became apparent that when merging gauges, we were actually ignoring the timestamp but, which meant that if you had an existing point for a gauge, say with timestamp X, and you merged in another point with a timestamp Y, the first point would be entirely dropped.

This manifested itself as a fun bug where, due to the bucket width and flush interval defaults, a gauge being reported every 10 seconds was only ever emitted every 30 seconds.

Solution

While gauges nominally have "last write wins" behavior, that logic only applies for a given timestamp. This PR fixes the logic of merging gauges such that we don't drop the previous points, and only apply the LWW behavior when merging two points with an identical timestamp.

Further, we've added a number of unit tests to ensure the merging behavior is correct for each metric type. We've additionally added a number of From and Display impls on the related types... some of which we only needed for initial debugging (but might be valuable in the future) and some of which are used by the unit tests.

[tobz]

This is the fix, basically. Right here. This one lil guy.

[tobz]

• chore: fold OriginKey hash into main hash to shrink ContextKey #452
• fix: merge gauges properly and avoid dropping existing points #444 👈 (View in Graphite)
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[pr-commenter]

Regression Detector (DogStatsD)

Regression Detector Results

Run ID: a37173bd-7754-4503-9ef7-af66b2e5a1c7

Baseline: 7.62.0-rc.8
Comparison: 7.62.0-rc.8

Optimization Goals: ✅ No significant changes detected

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
➖	dsd_uds_100mb_3k_contexts_distributions_only	memory utilization	+1.48	[+1.28, +1.69]	1
➖	quality_gates_idle_rss	memory utilization	+0.42	[+0.31, +0.52]	1
➖	dsd_uds_10mb_3k_contexts	ingress throughput	+0.00	[-0.00, +0.00]	1
➖	dsd_uds_1mb_3k_contexts	ingress throughput	+0.00	[-0.00, +0.00]	1
➖	dsd_uds_512kb_3k_contexts	ingress throughput	+0.00	[-0.01, +0.01]	1
➖	dsd_uds_1mb_50k_contexts	ingress throughput	+0.00	[-0.00, +0.00]	1
➖	dsd_uds_100mb_250k_contexts	ingress throughput	-0.00	[-0.00, +0.00]	1
➖	dsd_uds_1mb_3k_contexts_dualship	ingress throughput	-0.00	[-0.00, +0.00]	1
➖	dsd_uds_1mb_50k_contexts_memlimit	ingress throughput	-0.00	[-0.00, +0.00]	1
➖	dsd_uds_100mb_3k_contexts	ingress throughput	-0.00	[-0.05, +0.04]	1
➖	dsd_uds_40mb_12k_contexts_40_senders	ingress throughput	-0.00	[-0.00, +0.00]	1
➖	dsd_uds_500mb_3k_contexts	ingress throughput	-0.27	[-0.30, -0.24]	1

Bounds Checks: ❌ Failed

perf	experiment	bounds_check_name	replicates_passed	links
❌	quality_gates_idle_rss	memory_usage	0/10

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".

[pr-commenter]

Regression Detector (Saluki)

Regression Detector Results

Run ID: bb0a943d-79e1-47f1-9d95-58230bc95bbd

Baseline: 3576732
Comparison: 0d1856f
Diff

❌ Experiments with missing or malformed data

This is a critical error. No usable optimization goal data was produced by the listed experiments. This may be a result of misconfiguration. Ping #single-machine-performance and we can help out.

• dsd_uds_100mb_250k_contexts

Optimization Goals: ✅ No significant changes detected

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
➖	dsd_uds_500mb_3k_contexts	ingress throughput	+3.26	[+3.14, +3.39]	1
➖	dsd_uds_100mb_3k_contexts_distributions_only	memory utilization	+0.86	[+0.73, +0.99]	1
➖	dsd_uds_50mb_10k_contexts_no_inlining	ingress throughput	+0.02	[-0.05, +0.09]	1
➖	dsd_uds_50mb_10k_contexts_no_inlining_no_allocs	ingress throughput	+0.01	[-0.05, +0.07]	1
➖	dsd_uds_1mb_3k_contexts_dualship	ingress throughput	+0.00	[-0.00, +0.01]	1
➖	dsd_uds_40mb_12k_contexts_40_senders	ingress throughput	+0.00	[-0.02, +0.03]	1
➖	dsd_uds_100mb_3k_contexts	ingress throughput	+0.00	[-0.05, +0.05]	1
➖	dsd_uds_512kb_3k_contexts	ingress throughput	+0.00	[-0.01, +0.01]	1
➖	dsd_uds_1mb_50k_contexts	ingress throughput	-0.01	[-0.02, +0.01]	1
➖	dsd_uds_1mb_3k_contexts	ingress throughput	-0.01	[-0.03, +0.00]	1
➖	dsd_uds_10mb_3k_contexts	ingress throughput	-0.02	[-0.04, +0.01]	1
➖	quality_gates_idle_rss	memory utilization	-0.12	[-0.15, -0.09]	1
➖	dsd_uds_1mb_50k_contexts_memlimit	ingress throughput	-3.50	[-4.00, -3.01]	1

Bounds Checks: ✅ Passed

perf	experiment	bounds_check_name	replicates_passed	links
✅	quality_gates_idle_rss	memory_usage	10/10

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".

[pr-commenter]

Regression Detector Links

Experiment Result Links

experiment	link(s)
dsd_uds_100mb_250k_contexts	[Profiling (ADP)] [Profiling (DSD)] [SMP Dashboard]
dsd_uds_100mb_3k_contexts	[Profiling (ADP)] [Profiling (DSD)] [SMP Dashboard]
dsd_uds_100mb_3k_contexts_distributions_only	[Profiling (ADP)] [Profiling (DSD)] [SMP Dashboard]
dsd_uds_10mb_3k_contexts	[Profiling (ADP)] [Profiling (DSD)] [SMP Dashboard]
dsd_uds_1mb_3k_contexts	[Profiling (ADP)] [Profiling (DSD)] [SMP Dashboard]
dsd_uds_1mb_3k_contexts_dualship	[Profiling (ADP)] [Profiling (DSD)] [SMP Dashboard]
dsd_uds_1mb_50k_contexts	[Profiling (ADP)] [Profiling (DSD)] [SMP Dashboard]
dsd_uds_1mb_50k_contexts_memlimit	[Profiling (ADP)] [Profiling (DSD)] [SMP Dashboard]
dsd_uds_40mb_12k_contexts_40_senders	[Profiling (ADP)] [Profiling (DSD)] [SMP Dashboard]
dsd_uds_500mb_3k_contexts	[Profiling (ADP)] [Profiling (DSD)] [SMP Dashboard]
dsd_uds_512kb_3k_contexts	[Profiling (ADP)] [Profiling (DSD)] [SMP Dashboard]
quality_gates_idle_rss	[Profiling (ADP)] [Profiling (DSD)] [SMP Dashboard]
dsd_uds_50mb_10k_contexts_no_inlining (ADP only)	[Profiling (ADP)] [SMP Dashboard]
dsd_uds_50mb_10k_contexts_no_inlining_no_allocs (ADP only)	[Profiling (ADP)] [SMP Dashboard]

The base branch was changed.

The base branch was changed.