chore(agent-data-plane): update health registry worker to allow being restarted #1176

tobz · 2026-02-09T03:42:40Z

Summary

This PR slightly refactors HealthWorker and the underlying health registry runner code to support being able to restart the health registry worker.

Prior to this PR, spawning the health registry worker would fail un subsequent attempts since the receiver used to registry new components into the registry was already consumed by the first call to spawn the worker. We've simply added the ability to return the receiver and reset the state such that subsequent attempts to spawn the worker can take the receiver. We're still limited by only being able to have a single health registry worker at a time, but at least we can now cleanly recover from it being restarted.

Change Type

Bug fix
New feature
Non-functional (chore, refactoring, docs)
Performance

How did you test this PR?

Existing and new unit tests.

References

AGTMETRICS-393

tobz · 2026-02-09T03:43:04Z

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

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

Copilot

Pull request overview

Adds a shutdown-aware, restartable health registry runner so the registry can be spawned again after the runner stops.

Changes:

Update HealthRegistry::spawn to accept a shutdown future and allow respawning after the runner ends.
Introduce RunnerGuard to return the liveness response receiver back into registry state on shutdown.
Add tests for “duplicate spawn while running” and “respawn after shutdown”.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-02-09T03:44:41Z

lib/saluki-health/src/lib.rs

+            let mut inner = self.registry.lock().unwrap();
+            inner.responses_rx = Some(rx);
+            debug!("Returned response receiver to registry state.");


Drop should avoid panicking. Using lock().unwrap() inside drop() can panic if the mutex is poisoned, and a panic during drop can lead to aborts (especially if another panic is already in flight). Prefer handling poisoning explicitly (e.g., recover the inner value) and avoid unwrap() in drop().

Suggested change

let mut inner = self.registry.lock().unwrap();

inner.responses_rx = Some(rx);

debug!("Returned response receiver to registry state.");

match self.registry.lock() {

Ok(mut inner) => {

inner.responses_rx = Some(rx);

debug!("Returned response receiver to registry state.");

}

Err(poisoned) => {

let mut inner = poisoned.into_inner();

inner.responses_rx = Some(rx);

debug!("Returned response receiver to registry state after mutex poisoning.");

}

}

Copilot · 2026-02-09T03:44:41Z

lib/saluki-health/src/lib.rs

+        // Take the response receiver out of the guard so we can use it in the select loop.
+        // It will be put back when the guard is dropped.
+        let mut responses_rx = self
+            .guard
+            .responses_rx
+            .take()
+            .expect("responses_rx should always be Some when Runner is created");


Taking responses_rx out of RunnerGuard means the receiver will not be returned to the registry if the task is cancelled/aborted or unwinds before reaching the “put it back” code path. That breaks the documented goal of being restartable “after shutdown or an error”. Consider an RAII pattern that guarantees the receiver is put back even on early-exit (e.g., a small local guard whose Drop moves responses_rx back), or keep the receiver inside RunnerGuard and only borrow it mutably for recv().

Copilot · 2026-02-09T03:44:41Z

lib/saluki-health/src/lib.rs

+        };
+
+        for component_id in 0..component_count {
+            self.process_component_health_update(component_id, HealthUpdate::Unknown);


On runner (re)start, this forces every existing component’s health to Unknown, which can erase the last-known state and potentially emit unnecessary state transitions/notifications. If the intent is to “pick up where it left off”, consider scheduling immediate probes without overwriting existing health, or only setting Unknown for components that truly have no prior health value.

Suggested change

self.process_component_health_update(component_id, HealthUpdate::Unknown);

// Do not overwrite existing health with `Unknown` on (re)start; just schedule immediate probes.

pr-commenter · 2026-02-09T03:53:02Z

Binary Size Analysis (Agent Data Plane)

Target: 9fe09ef (baseline) vs 7ca8cbc (comparison) diff
Analysis Type: Stripped binaries (debug symbols excluded)
Baseline Size: 26.96 MiB
Comparison Size: 27.10 MiB
Size Change: +146.62 KiB (+0.53%)
Pass/Fail Threshold: +5%
Result: PASSED ✅

Changes by Module

Module	File Size	Symbols
`saluki_core::runtime::supervisor`	+68.92 KiB	59
`core`	+43.71 KiB	11166
`tokio`	+24.10 KiB	2717
`agent_data_plane::internal::initialize_and_launch_runtime`	-22.07 KiB	2
`std`	-11.37 KiB	280
`[sections]`	+10.34 KiB	10
`saluki_core::runtime::process`	+9.33 KiB	7
`agent_data_plane::internal::control_plane`	-8.03 KiB	26
`anyhow`	+6.94 KiB	1286
`saluki_core::runtime::dedicated`	+5.63 KiB	4
`agent_data_plane::internal::observability`	+5.60 KiB	16
`saluki_core::topology::running`	+5.50 KiB	31
`[Unmapped]`	-5.18 KiB	1
`saluki_app::metrics::collect_runtime_metrics`	-4.73 KiB	1
`saluki_core::runtime::restart`	+3.50 KiB	7
`saluki_core::runtime::shutdown`	+2.27 KiB	4
`tracing_core`	+2.24 KiB	521
`alloc`	+2.10 KiB	1222
`saluki_health::Runner::run`	+1.89 KiB	8
`hashbrown`	+1.81 KiB	412

Detailed Symbol Changes

    FILE SIZE        VM SIZE    
 --------------  -------------- 
  [NEW] +1.79Mi  [NEW] +1.79Mi    std::thread::local::LocalKey<T>::with::h938f6ddf7aedbbb8
  +1.1%  +170Ki  +1.0%  +141Ki    [29783 Others]
  [NEW]  +113Ki  [NEW]  +113Ki    agent_data_plane::cli::run::create_topology::_{{closure}}::hfb0b645d0bb8c8cf
  [NEW] +68.2Ki  [NEW] +68.1Ki    h2::hpack::decoder::Decoder::try_decode_string::hd1c9a39c48e78a6e
  [NEW] +63.7Ki  [NEW] +63.6Ki    saluki_components::common::datadog::io::run_endpoint_io_loop::_{{closure}}::hd35493454b2aefa0
  [NEW] +63.5Ki  [NEW] +63.2Ki    _<agent_data_plane::internal::control_plane::PrivilegedApiWorker as saluki_core::runtime::supervisor::Supervisable>::initialize::_{{closure}}::hc40718d721e65675
  [NEW] +62.3Ki  [NEW] +62.2Ki    agent_data_plane::main::_{{closure}}::hd35953e90113f5c5
  [NEW] +55.7Ki  [NEW] +55.6Ki    agent_data_plane::cli::run::handle_run_command::_{{closure}}::hb1b79147c6e6e2ef
  [NEW] +48.8Ki  [NEW] +48.7Ki    saluki_app::bootstrap::AppBootstrapper::bootstrap::_{{closure}}::hc8e634acdf97ad33
  [NEW] +47.7Ki  [NEW] +47.6Ki    moka::sync::base_cache::Inner<K,V,S>::do_run_pending_tasks::h0ef6755dbc77ea2a
  [NEW] +46.1Ki  [NEW] +46.0Ki    h2::proto::connection::Connection<T,P,B>::poll::h1671860da0918c66
  [DEL] -46.1Ki  [DEL] -46.0Ki    h2::proto::connection::Connection<T,P,B>::poll::h2aedcbe1089b311c
  [DEL] -47.7Ki  [DEL] -47.6Ki    moka::sync::base_cache::Inner<K,V,S>::do_run_pending_tasks::ha97a2f55834f17a3
  [DEL] -48.8Ki  [DEL] -48.7Ki    saluki_app::bootstrap::AppBootstrapper::bootstrap::_{{closure}}::hbb5d0d8e944d3a21
  [DEL] -57.8Ki  [DEL] -57.7Ki    agent_data_plane::cli::run::handle_run_command::_{{closure}}::hd8d13580d16cc8a3
  [DEL] -62.3Ki  [DEL] -62.2Ki    agent_data_plane::main::_{{closure}}::h1c7c640002ce8ad1
  [DEL] -63.8Ki  [DEL] -63.6Ki    saluki_components::common::datadog::io::run_endpoint_io_loop::_{{closure}}::hc194831d658db8cc
  [DEL] -68.2Ki  [DEL] -68.1Ki    h2::hpack::decoder::Decoder::try_decode_string::hbfc0bb5a77e7669f
  [DEL] -84.5Ki  [DEL] -84.4Ki    agent_data_plane::internal::control_plane::spawn_control_plane::_{{closure}}::heace61ab0f2bde0c
  [DEL]  -113Ki  [DEL]  -113Ki    agent_data_plane::cli::run::create_topology::_{{closure}}::h6a8b41fa76a89e2c
  [DEL] -1.79Mi  [DEL] -1.79Mi    std::thread::local::LocalKey<T>::with::h0d8770940d38805b
  +0.5%  +146Ki  +0.5%  +117Ki    TOTAL

pr-commenter · 2026-02-09T04:23:43Z

Regression Detector (Agent Data Plane)

Regression Detector Results

Run ID: ff31bc77-4d94-4ffb-8380-73b962e96ea9

Baseline: 9fe09ef
Comparison: 7ca8cbc
Diff

❌ Experiments with retried target crashes

This is a critical error. One or more replicates failed with a non-zero exit code. These replicates may have been retried. See Replicate Execution Details for more information.

quality_gates_rss_dsd_ultraheavy
otlp_ingest_logs_5mb_cpu
otlp_ingest_traces_5mb_cpu
otlp_ingest_logs_5mb_memory
quality_gates_rss_dsd_medium

Optimization Goals: ✅ No significant changes detected

Experiments ignored for regressions

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

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
❌	otlp_ingest_logs_5mb_memory	memory utilization	+10.71	[+10.00, +11.43]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_cpu	% cpu utilization	+0.47	[-4.88, +5.82]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_throughput	ingress throughput	+0.03	[-0.10, +0.16]	1	(metrics) (profiles) (logs)

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
❌	otlp_ingest_logs_5mb_memory	memory utilization	+10.71	[+10.00, +11.43]	1	(metrics) (profiles) (logs)
➖	dsd_uds_512kb_3k_contexts_cpu	% cpu utilization	+4.87	[-50.63, +60.37]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_cpu	% cpu utilization	+4.43	[-47.32, +56.18]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_idle	memory utilization	+3.01	[+2.94, +3.08]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_cpu	% cpu utilization	+2.17	[-3.85, +8.19]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_memory	memory utilization	+2.15	[+1.96, +2.35]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_memory	memory utilization	+1.74	[+1.55, +1.93]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_medium	memory utilization	+1.38	[+1.19, +1.57]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_5mb_cpu	% cpu utilization	+1.33	[-1.27, +3.93]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_memory	memory utilization	+1.20	[+1.01, +1.38]	1	(metrics) (profiles) (logs)
➖	dsd_uds_100mb_3k_contexts_memory	memory utilization	+1.11	[+0.92, +1.30]	1	(metrics) (profiles) (logs)
➖	dsd_uds_512kb_3k_contexts_memory	memory utilization	+0.92	[+0.74, +1.10]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_low	memory utilization	+0.89	[+0.71, +1.06]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_cpu	% cpu utilization	+0.47	[-4.88, +5.82]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_heavy	memory utilization	+0.45	[+0.33, +0.58]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_ultraheavy	memory utilization	+0.28	[+0.15, +0.41]	1	(metrics) (profiles) (logs)
➖	dsd_uds_100mb_3k_contexts_cpu	% cpu utilization	+0.25	[-5.75, +6.24]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_cpu	% cpu utilization	+0.03	[-1.24, +1.31]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_throughput	ingress throughput	+0.03	[-0.10, +0.16]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_throughput	ingress throughput	+0.02	[-0.10, +0.14]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_throughput	ingress throughput	+0.01	[-0.15, +0.17]	1	(metrics) (profiles) (logs)
➖	dsd_uds_512kb_3k_contexts_throughput	ingress throughput	+0.00	[-0.05, +0.05]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_throughput	ingress throughput	-0.00	[-0.06, +0.06]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_5mb_throughput	ingress throughput	-0.01	[-0.09, +0.07]	1	(metrics) (profiles) (logs)
➖	dsd_uds_100mb_3k_contexts_throughput	ingress throughput	-0.03	[-0.08, +0.03]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_throughput	ingress throughput	-0.11	[-0.23, +0.02]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_5mb_memory	memory utilization	-0.40	[-0.65, -0.15]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_memory	memory utilization	-1.64	[-1.88, -1.41]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_cpu	% cpu utilization	-4.97	[-34.86, +24.92]	1	(metrics) (profiles) (logs)

Bounds Checks: ✅ Passed

perf	experiment	bounds_check_name	replicates_passed	links
✅	quality_gates_rss_dsd_heavy	memory_usage	10/10	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_low	memory_usage	10/10	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_medium	memory_usage	10/10	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_ultraheavy	memory_usage	10/10	(metrics) (profiles) (logs)
✅	quality_gates_rss_idle	memory_usage	10/10	(metrics) (profiles) (logs)

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:

Its estimated |Δ mean %| ≥ 5.00%, indicating the change is big enough to merit a closer look.
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.
Its configuration does not mark it "erratic".

Replicate Execution Details

We run multiple replicates for each experiment/variant. However, we allow replicates to be automatically retried if there are any failures, up to 8 times, at which point the replicate is marked dead and we are unable to run analysis for the entire experiment. We call each of these attempts at running replicates a replicate execution. This section lists all replicate executions that failed due to the target crashing or being oom killed.

Note: In the below tables we bucket failures by experiment, variant, and failure type. For each of these buckets we list out the replicate indexes that failed with an annotation signifying how many times said replicate failed with the given failure mode. In the below example the baseline variant of the experiment named experiment_with_failures had two replicates that failed by oom kills. Replicate 0, which failed 8 executions, and replicate 1 which failed 6 executions, all with the same failure mode.

Experiment	Variant	Replicates	Failure	Logs	Debug Dashboard
experiment_with_failures	baseline	0 (x8) 1 (x6)	Oom killed		Debug Dashboard

The debug dashboard links will take you to a debugging dashboard specifically designed to investigate replicate execution failures.

❌ Retried Normal Replicate Execution Failures (non-profiling)

Experiment	Variant	Replicates	Failure	Debug Dashboard
otlp_ingest_logs_5mb_cpu	comparison	0	Failed to shutdown when requested	Debug Dashboard
otlp_ingest_logs_5mb_memory	comparison	2, 1	Failed to shutdown when requested	Debug Dashboard
otlp_ingest_traces_5mb_cpu	baseline	1	Failed to shutdown when requested	Debug Dashboard
quality_gates_rss_dsd_medium	baseline	5	Failed to shutdown when requested	Debug Dashboard
quality_gates_rss_dsd_ultraheavy	baseline	6	Failed to shutdown when requested	Debug Dashboard

…stry

Copilot AI review requested due to automatic review settings February 9, 2026 03:42

dd-octo-sts bot added the area/core Core functionality, event model, etc. label Feb 9, 2026

This was referenced Feb 9, 2026

enhancement(core): add support for spawning supervisors on a dedicated runtime #1145

Open

enhancement(core): make initialization async in Supervisable #1146

Open

Copilot AI reviewed Feb 9, 2026

View reviewed changes

tobz added the type/chore Updates to dependencies or general "administrative" tasks necessary to maintain the codebase/repo. label Feb 9, 2026

tobz changed the title ~~enhancement(agent-data-plane): add restartable worker for health registry~~ chore(agent-data-plane): add restartable worker for health registry Feb 9, 2026

tobz changed the title ~~chore(agent-data-plane): add restartable worker for health registry~~ chore(agent-data-plane): update health registry worker to allow being restarted Feb 9, 2026

tobz marked this pull request as ready for review February 9, 2026 14:12

tobz requested a review from a team as a code owner February 9, 2026 14:12

enhancement(agent-data-plane): add restartable worker for health regi…

7ca8cbc

…stry

tobz force-pushed the tobz/adp-move-int-o11y-control-plane-to-supervisor branch from 70c54c9 to b72ece8 Compare February 11, 2026 16:24

tobz force-pushed the tobz/supervisor-health-registry-worker branch from a0e1d19 to 7ca8cbc Compare February 11, 2026 16:24

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

chore(agent-data-plane): update health registry worker to allow being restarted #1176

chore(agent-data-plane): update health registry worker to allow being restarted #1176

tobz commented Feb 9, 2026 •

edited

Loading

Uh oh!

tobz commented Feb 9, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Feb 9, 2026

Uh oh!

Copilot AI Feb 9, 2026

Uh oh!

Copilot AI Feb 9, 2026

Uh oh!

pr-commenter bot commented Feb 9, 2026 •

edited

Loading

Changes by Module

Detailed Symbol Changes

Uh oh!

pr-commenter bot commented Feb 9, 2026 •

edited

Loading

Experiments ignored for regressions

Fine details of change detection per experiment

Bounds Checks: ✅ Passed

Explanation

Replicate Execution Details

❌ Retried Normal Replicate Execution Failures (non-profiling)

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

	self.process_component_health_update(component_id, HealthUpdate::Unknown);
	// Do not overwrite existing health with `Unknown` on (re)start; just schedule immediate probes.

chore(agent-data-plane): update health registry worker to allow being restarted #1176

Are you sure you want to change the base?

chore(agent-data-plane): update health registry worker to allow being restarted #1176

Conversation

tobz commented Feb 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Change Type

How did you test this PR?

References

Uh oh!

tobz commented Feb 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

Uh oh!

pr-commenter bot commented Feb 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Binary Size Analysis (Agent Data Plane)

Changes by Module

Detailed Symbol Changes

Uh oh!

pr-commenter bot commented Feb 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Regression Detector (Agent Data Plane)

Regression Detector Results

❌ Experiments with retried target crashes

Optimization Goals: ✅ No significant changes detected

Experiments ignored for regressions

Fine details of change detection per experiment

Bounds Checks: ✅ Passed

Explanation

Replicate Execution Details

❌ Retried Normal Replicate Execution Failures (non-profiling)

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

tobz commented Feb 9, 2026 •

edited

Loading

tobz commented Feb 9, 2026 •

edited

Loading

pr-commenter bot commented Feb 9, 2026 •

edited

Loading

pr-commenter bot commented Feb 9, 2026 •

edited

Loading