[Backport 7.80.x] fix(npm/windows): suppress lingering flow re-reports that cause >100% TCP failure rate

[dd-octo-sts]

Backport 4d8c72b from #50823.

---

What does this PR do?

On Windows, the ddnpm kernel driver can leave a terminated flow stuck in openFlows when a WFP transport callout holds a refcount > 2 at the moment the closure callback fires. Each agent poll re-reports the flow with a fresh TCPFailures={104:1}. Since TCPFailures is not tracked in StatCounters, it has no delta baseline and ships as a new failure every poll while Last.TCPClosed stays at 0 after the first poll — causing the backend rate to climb past 100% unboundedly.

Fix: in updateConnWithStats, when a connection was already reported as closed (sts.TCPClosed > 0) and the current delta has no new close event (last.TCPClosed == 0), zero out TCPFailures so the re-report is empty and filtered from the payload entirely.

Adds a windows_lingering_flows telemetry counter and warn-level log line in system-probe.log so affected hosts are observable.

Motivation

https://datadoghq.atlassian.net/browse/WINA-2711

Describe how you validated your changes

Unit test in pkg/network/state_tcp_failure_replay_test.go reproduces the state-machine asymmetry across 3 polls and verifies the fix holds the backend rate at exactly 100%, with the windowsLingeringFlows counter incrementing once per suppressed re-report.

Reproduced end-to-end on EC2 (Windows Server 2025, real ENI) with ~10,000 TLS+HTTP RST connections at 256 workers. Before the fix: NPM rate > 100% and climbing. After deploying the patched system-probe.exe: system-probe.log showed [74 Windows lingering flows suppressed] and the NPM rate held at 100%.

Additional Notes

The permanent fix belongs in the ddnpm driver (flowdata.c — the InterlockedCompareExchange in DecrementRefCountAndMoveFromOpenToClosedTable should retry or use a different sync primitive). This is an agent-side workaround to prevent the symptom from reaching customers while the driver fix is developed.

Note: loopback traffic is synchronous in the Windows kernel and cannot trigger the race. Real NIC traffic is required to reproduce.

[datadog-prod-us1-4]

✨ Fix all issues with BitsAI

⚠️ Warnings

🚦 1 Pipeline job failed

DataDog/datadog-agent | oracle: [21.3.0-xe]     

🛟 This job is unlikely to succeed on retry. Please review your pipeline configuration.

Failed to ping Oracle instance: ORA-12514: TNS:listener does not currently know of service requested in connect descriptor.

ℹ️ Info

🎯 Code Coverage (details)
• Patch Coverage: 100.00%
• Overall Coverage: 50.31% (-0.00%)

Useful? React with 👍 / 👎

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 85a06eb | Docs | Datadog PR Page | Give us feedback!}

[dd-octo-sts]

Files inventory check summary

File checks results against ancestor 9558f092:

Results for datadog-agent_7.80.0~rc.2.git.4.85a06eb.pipeline.114385449-1_amd64.deb:

No change detected

[dd-octo-sts]

Static quality checks

✅ Please find below the results from static quality gates
Comparison made with ancestor 9558f09
📊 Static Quality Gates Dashboard
🔗 SQG Job

Successful checks

Info

	Quality gate	Change	Size (prev → curr → max)
✅	agent_msi	+2.0 KiB (0.00% increase)	609.584 → 609.586 → 624.040
✅	docker_agent_jmx_arm64	+2.92 KiB (0.00% increase)	986.644 → 986.647 → 989.800

31 successful checks with minimal change (< 2 KiB)

	Quality gate	Current Size
✅	agent_deb_amd64	744.381 MiB
✅	agent_deb_amd64_fips	702.393 MiB
✅	agent_heroku_amd64	310.088 MiB
✅	agent_rpm_amd64	744.364 MiB
✅	agent_rpm_amd64_fips	702.376 MiB
✅	agent_rpm_arm64	721.970 MiB
✅	agent_rpm_arm64_fips	683.124 MiB
✅	agent_suse_amd64	744.364 MiB
✅	agent_suse_amd64_fips	702.376 MiB
✅	agent_suse_arm64	721.970 MiB
✅	agent_suse_arm64_fips	683.124 MiB
✅	docker_agent_amd64	804.522 MiB
✅	docker_agent_arm64	806.948 MiB
✅	docker_agent_jmx_amd64	995.441 MiB
✅	docker_cluster_agent_amd64	206.858 MiB
✅	docker_cluster_agent_arm64	220.826 MiB
✅	docker_cws_instrumentation_amd64	7.154 MiB
✅	docker_cws_instrumentation_arm64	6.689 MiB
✅	docker_dogstatsd_amd64	39.503 MiB
✅	docker_dogstatsd_arm64	37.690 MiB
✅	docker_host_profiler_amd64	302.254 MiB
✅	docker_host_profiler_arm64	313.761 MiB
✅	dogstatsd_deb_amd64	30.162 MiB
✅	dogstatsd_deb_arm64	28.284 MiB
✅	dogstatsd_rpm_amd64	30.162 MiB
✅	dogstatsd_suse_amd64	30.162 MiB
✅	iot_agent_deb_amd64	44.457 MiB
✅	iot_agent_deb_arm64	41.421 MiB
✅	iot_agent_deb_armhf	42.130 MiB
✅	iot_agent_rpm_amd64	44.457 MiB
✅	iot_agent_suse_amd64	44.457 MiB

[cit-pr-commenter-54b7da]

Regression Detector

Regression Detector Results

Metrics dashboard
Target profiles
Run ID: 78d50896-ac87-4cc2-84d0-4285972ca88e

Baseline: 9558f09
Comparison: 85a06eb
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_gate_idle

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
➖	docker_containers_cpu	% cpu utilization	+1.49	[-1.45, +4.43]	1	Logs

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
➖	docker_containers_cpu	% cpu utilization	+1.49	[-1.45, +4.43]	1	Logs
➖	quality_gate_metrics_logs	memory utilization	+0.51	[+0.27, +0.76]	1	Logs bounds checks dashboard
➖	tcp_syslog_to_blackhole	ingress throughput	+0.33	[+0.12, +0.54]	1	Logs
➖	ddot_logs	memory utilization	+0.32	[+0.26, +0.37]	1	Logs
➖	ddot_metrics_sum_cumulativetodelta_exporter	memory utilization	+0.21	[-0.02, +0.45]	1	Logs
➖	docker_containers_memory	memory utilization	+0.20	[+0.10, +0.30]	1	Logs
➖	quality_gate_logs	% cpu utilization	+0.19	[-0.80, +1.18]	1	Logs bounds checks dashboard
➖	quality_gate_idle	memory utilization	+0.16	[+0.11, +0.20]	1	Logs bounds checks dashboard
➖	ddot_metrics_sum_cumulative	memory utilization	+0.15	[-0.00, +0.31]	1	Logs
➖	uds_dogstatsd_20mb_12k_contexts_20_senders	memory utilization	+0.15	[+0.10, +0.20]	1	Logs
➖	file_to_blackhole_1000ms_latency	egress throughput	+0.03	[-0.42, +0.48]	1	Logs
➖	uds_dogstatsd_to_api	ingress throughput	+0.01	[-0.21, +0.23]	1	Logs
➖	uds_dogstatsd_to_api_v3	ingress throughput	-0.00	[-0.20, +0.19]	1	Logs
➖	tcp_dd_logs_filter_exclude	ingress throughput	-0.01	[-0.10, +0.08]	1	Logs
➖	file_to_blackhole_0ms_latency	egress throughput	-0.02	[-0.55, +0.51]	1	Logs
➖	quality_gate_idle_all_features	memory utilization	-0.02	[-0.06, +0.02]	1	Logs bounds checks dashboard
➖	file_tree	memory utilization	-0.02	[-0.07, +0.04]	1	Logs
➖	file_to_blackhole_100ms_latency	egress throughput	-0.07	[-0.25, +0.10]	1	Logs
➖	file_to_blackhole_500ms_latency	egress throughput	-0.08	[-0.48, +0.31]	1	Logs
➖	ddot_metrics_sum_delta	memory utilization	-0.13	[-0.32, +0.07]	1	Logs
➖	otlp_ingest_logs	memory utilization	-0.35	[-0.46, -0.24]	1	Logs
➖	otlp_ingest_metrics	memory utilization	-0.41	[-0.57, -0.25]	1	Logs
➖	ddot_metrics	memory utilization	-0.60	[-0.80, -0.39]	1	Logs

Bounds Checks: ✅ Passed

perf	experiment	bounds_check_name	replicates_passed	observed_value	links
✅	docker_containers_cpu	simple_check_run	10/10	703 ≥ 26
✅	docker_containers_memory	memory_usage	10/10	245.53MiB ≤ 370MiB
✅	docker_containers_memory	simple_check_run	10/10	719 ≥ 26
✅	file_to_blackhole_0ms_latency	memory_usage	10/10	0.17GiB ≤ 1.20GiB
✅	file_to_blackhole_0ms_latency	missed_bytes	10/10	0B = 0B
✅	file_to_blackhole_1000ms_latency	memory_usage	10/10	0.20GiB ≤ 1.20GiB
✅	file_to_blackhole_1000ms_latency	missed_bytes	10/10	0B = 0B
✅	file_to_blackhole_100ms_latency	memory_usage	10/10	0.17GiB ≤ 1.20GiB
✅	file_to_blackhole_100ms_latency	missed_bytes	10/10	0B = 0B
✅	file_to_blackhole_500ms_latency	memory_usage	10/10	0.18GiB ≤ 1.20GiB
✅	file_to_blackhole_500ms_latency	missed_bytes	10/10	0B = 0B
✅	quality_gate_idle	intake_connections	10/10	3 ≤ 4	bounds checks dashboard
✅	quality_gate_idle	memory_usage	10/10	142.76MiB ≤ 147MiB	bounds checks dashboard
✅	quality_gate_idle_all_features	intake_connections	10/10	3 ≤ 4	bounds checks dashboard
✅	quality_gate_idle_all_features	memory_usage	10/10	477.66MiB ≤ 495MiB	bounds checks dashboard
✅	quality_gate_logs	intake_connections	10/10	3 ≤ 6	bounds checks dashboard
✅	quality_gate_logs	memory_usage	10/10	175.57MiB ≤ 195MiB	bounds checks dashboard
✅	quality_gate_logs	missed_bytes	10/10	0B = 0B	bounds checks dashboard
✅	quality_gate_metrics_logs	cpu_usage	10/10	365.06 ≤ 2000	bounds checks dashboard
✅	quality_gate_metrics_logs	intake_connections	10/10	3 ≤ 6	bounds checks dashboard
✅	quality_gate_metrics_logs	memory_usage	10/10	404.21MiB ≤ 430MiB	bounds checks dashboard
✅	quality_gate_metrics_logs	missed_bytes	10/10	0B = 0B	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".

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
quality_gate_idle	baseline	9	Oom killed	Debug Dashboard

CI Pass/Fail Decision

✅ Passed. All Quality Gates passed.

• quality_gate_idle_all_features, bounds check intake_connections: 10/10 replicas passed. Gate passed.
• quality_gate_idle_all_features, bounds check memory_usage: 10/10 replicas passed. Gate passed.
• quality_gate_metrics_logs, bounds check memory_usage: 10/10 replicas passed. Gate passed.
• quality_gate_metrics_logs, bounds check missed_bytes: 10/10 replicas passed. Gate passed.
• quality_gate_metrics_logs, bounds check intake_connections: 10/10 replicas passed. Gate passed.
• quality_gate_metrics_logs, bounds check cpu_usage: 10/10 replicas passed. Gate passed.
• quality_gate_idle, bounds check memory_usage: 10/10 replicas passed. Gate passed.
• quality_gate_idle, bounds check intake_connections: 10/10 replicas passed. Gate passed.
• quality_gate_logs, bounds check missed_bytes: 10/10 replicas passed. Gate passed.
• quality_gate_logs, bounds check memory_usage: 10/10 replicas passed. Gate passed.
• quality_gate_logs, bounds check intake_connections: 10/10 replicas passed. Gate passed.