Ettore Di Giacinto d1a59aae41 fix(failover): never load a warm target inside a probe ...

A warm target's liveness probe called ModelLoader.Load, which blocked
until the model finished loading (while the warm preload loaded it
too). Tick waited for every probe, so all probing froze, and the probe
then ran HealthCheck on an expired context and tripped the target at
every startup.

The prober now takes a function that returns the running backend
without loading it. A target that is not loaded passes liveness; its
recovery is neither confirmed nor failed and it returns to healthy
after min_dwell, like a cold target. Tick no longer waits for probes:
each probe applies its own result and a target whose probe is running
is skipped.

Assisted-by: Claude:claude-opus-5-5

2026-09-26 19:22:14 +00:00

..

agent_jobs.go

feat(distributed): SyncedMap component + migrate finetune/quant/agent-tasks to cross-replica state (#10542)

2026-06-27 23:23:51 +02:00

application_suite_test.go

feat(branding): admin-configurable instance name, tagline, and assets (#9635)

2026-05-02 15:51:36 +02:00

application.go

feat(failover): add MCP tools to list chains and pin targets

2026-09-26 18:18:46 +00:00

config_file_watcher.go

fix(gallery): verification follow-ups for oci:// galleries (#12243)

2026-09-24 21:20:21 +02:00

distributed.go

fix(distributed): exclude pinned models from cluster eviction and idle scale-down

2026-09-11 23:08:26 +00:00

failover_test.go

fix(failover): never load a warm target inside a probe

2026-09-26 19:22:14 +00:00

failover.go

fix(failover): never load a warm target inside a probe

2026-09-26 19:22:14 +00:00

mitm_test.go

fix(router): production-ready request router + auto-size batch for embedding/rerank (#10104)

2026-06-12 16:21:15 +02:00

mitm.go

feat(pii): NER tier engine — privacy-filter.cpp backend + NER-centric PII filter (#10360)

2026-06-18 11:45:22 +01:00

model_artifact_materializer_test.go

feat: materialize Hugging Face model artifacts (#10825)

2026-07-15 01:09:33 +02:00

p2p.go

fix(p2p): serialize access to p2pCtx/p2pCancel (#10839) (#10861)

2026-07-17 09:02:05 +02:00

pii_policy_test.go

feat(pii): NER tier engine — privacy-filter.cpp backend + NER-centric PII filter (#10360)

2026-06-18 11:45:22 +01:00

readiness_test.go

fix(http): make /readyz reflect startup readiness, plus gitignore and coverage-ratchet fixes (#10989)

2026-07-20 21:45:27 +02:00

router_factories_test.go

feat(router): make KNN a first-class classifier with a persisted, curated corpus (#10652)

2026-08-18 09:37:43 +02:00

router_factories.go

feat(router): make KNN a first-class classifier with a persisted, curated corpus (#10652)

2026-08-18 09:37:43 +02:00

runtime_settings_branding_test.go

refactor(settings): single declarative registry for runtime settings (fixes the #10845 bug class) (#10864)

2026-07-16 22:39:59 +02:00

runtime_settings_galleries_test.go

fix(gallery): verification follow-ups for oci:// galleries (#12243)

2026-09-24 21:20:21 +02:00

startup.go

fix(failover): never load a warm target inside a probe

2026-09-26 19:22:14 +00:00

upgrade_checker.go

fix(distributed): make per-node backend upgrade actually upgrade (#10838)

2026-07-15 09:16:55 +02:00

watchdog_test.go

fix(watchdog): guard StopWatchdog with watchdogMutex to prevent double close (#10841) (#10859)

2026-07-16 09:40:29 +02:00

watchdog.go

feat(failover): run the chain manager and keep warm targets loaded

2026-09-26 16:06:15 +00:00