[maxmad64bis]

⚠️ base-red inherited: #15306

Summary

Opt-in, default unchanged: with PROXY_HEALTH_PASSIVE_SKIP unset (default false) the sweep probes every proxy exactly as before. When enabled, the background sweep stops re-probing an endpoint that recent single-provider production successes already prove alive, so it sends fewer probes. Only an attributed proxy failure keeps the live probe, so confirmed failures still feed auto-disable and auto-remove.

Related Issues

• Related to chore(proxyHealth): rebase #14805 on #14776 + run proxy-health sweep suites (follow-up #14776) #14852 (rebase onto fix(proxyHealth): promote connection failures with cross-proxy evidence #14776, proxy-health suites on tip and after rebase, decision.ts kept intact — all addressed)

Validation

• Change type: routing
• Focused tests and category gates from the golden path
• npm run lint
• Reconciled with the current active release base; focused checks rerun afterward
• Production-code changes include a new or updated automated test in this PR

Tests Added Or Updated

• tests/unit/proxy-health-passive-verdict.test.ts: veto on any attributed failure in either order, rewritten from the inverted order.
• tests/unit/proxy-health-passive-sweep.test.ts (7 cases): healthy skip, kept probe, flag off, multi-provider guard, counted apart.
• tests/unit/proxy-health-passive-traffic.test.ts (2 mixed cases): real logger → storage → sweep, [fail, ok] and [ok, fail] both keep the probe.

Coverage Notes

• src/lib/proxyHealth/passiveVerdict.ts: proxy-health-passive-verdict.test.ts, plus the real logger path in proxy-health-passive-traffic.test.ts.
• src/lib/proxyHealth/scheduler.ts: proxy-health-passive-sweep.test.ts (skip, kept probe, flag off, promotion beside a skip).

Reviewer Notes

• Behavior: Skips are scheduling-only: a skipped endpoint never reaches the decision phase, so no streak, status, evidence or verdict display is written for it. Grouped by host and port: two entries with the same host and port but different ids share one verdict. TARGET_REFUSAL_STATUSES is a frozen ReadonlySet([401, 403, 429]) read with .has() only: no bound or size test needed. Position on fix(proxies): set aside egress after repeated transport failures with cross-egress success #14802 (gated flag-off): with the flag off the sweep reads nothing extra and probes as before.
• Merge: This PR shares .env.example and docs/reference/ENVIRONMENT.md (append-only additions at distinct anchors) with feat(proxy-subscription): verify the generated core configuration before replacing the file #15142, feat(proxy-subscription): reload the proxy core after a verified config replacement #15238, feat(proxy-subscription): prune nodes the core check rejects, then retry verification #15295 and feat(proxies): accept operator-provided dated egress observations for pool members #15314: merge order with those PRs is anchor-level only, no overlapping lines. With feat(i18n): forward the optional reasoning control in the translation backend #15462 (shared src/lib/proxyHealth/scheduler.ts): either order. With feat(proxy-subscription): warn with a translated count when generation skips entries #15564 (stacked proxy-core config chain on the same base files): either order. Disjoint from fix(dashboard): report headers-wait and client aborts outside transport failures #15248 (this PR: sweep scheduling skip; fix(dashboard): report headers-wait and client aborts outside transport failures #15248: health display counters) — either order; "abandoned" here means an abandoned send excluded from the skip proof, not the headers-wait/client-abort counters of fix(dashboard): report headers-wait and client aborts outside transport failures #15248.

[diegosouzapw]

Thanks — using production traffic to skip redundant probes is a sensible idea and the pure passiveVerdict module is easy to test. Three requests: (1) with recent traffic this changes default sweep behavior, so please put it behind an opt-in flag (or call it out explicitly); (2) TLS/DNS/fetch failed are broad enough to blame the provider or network on the proxy — can we tighten them? (3) it conflicts with #14776 in scheduler.ts, so please rebase once that one lands.

[diegosouzapw]

Heads-up: #14776 landed on release/v3.8.51 and also rewrote src/lib/proxyHealth/scheduler.ts, so this PR now conflicts. It needs a rebase onto the tip that keeps both behaviors. Tracking in #14852.

[maxmad64bis]

#14852 Done — all three requests addressed: opt-in flag PROXY_HEALTH_PASSIVE_SKIP (off by default), tightened attribution to bounded syscall codes, and rebase onto the tip keeping both sweep behaviors (scheduler.ts conflict-free, proxy-health* suites green on tip and after).

[diegosouzapw]

Re-homed to release/v3.8.52: v3.8.51 entered its release freeze, so the branch now belongs to the release captain and development continues on the next cycle. Nothing is wrong with this PR — it just needed a live base. No action needed from you; CI will re-run against the new base.