docs: record isolated release slot drill

2026-08-30 14:10:18 +03:00
parent 216d76f1d8
commit 5593f2ac2c
1 changed files with 51 additions and 0 deletions
@@ -0,0 +1,51 @@
+{
+  "task": "GEN2-DEPLOY-003",
+  "status": "PASS",
+  "timestamp_utc": "2026-08-30T11:09:35Z",
+  "root_cause": {
+    "stable_and_candidate_project": "assistant",
+    "stable_and_candidate_working_directory": "/home/m/assistant-deploy/repo/deploy",
+    "updater": "/home/m/assistant-deploy/update.sh",
+    "updater_schedule": "cron every five minutes",
+    "unsafe_operation": "the updater ran full docker compose up for the stable project and invoked the old in-project candidate canary"
+  },
+  "remediation": {
+    "stable_project": "assistant",
+    "candidate_project": "assistant-slot-b",
+    "candidate_file": "deploy/docker-compose.candidate-slot.yml",
+    "shared_resources": ["assistant_default external network", "llm-local_default external network", "existing secret and research volumes"],
+    "updater_ownership": "stable-only updater; candidate release command is not invoked by the updater",
+    "supervisor_release_set_slot": "ReleaseSet.slot persisted in durable Supervisor state"
+  },
+  "candidate_revision": "216d76f1d8b4eeba6ee9427da4c87c3e5c728c02",
+  "stable_revision": "2ee5da81145cdad2da1c92c24a91145ad4af61a5",
+  "candidate_components": {
+    "creator-web": "healthy",
+    "core-runtime": "healthy",
+    "cognitive-worker": "healthy",
+    "research-runtime": "healthy"
+  },
+  "drill": {
+    "stable_ids_captured_before_candidate": true,
+    "stable_ids_unchanged_after_candidate_start": true,
+    "header_canary": "candidate slot b",
+    "forced_failure": "candidate cognitive-worker stopped",
+    "rollback": "Supervisor release rollback returned traffic to legacy stable without recreating stable containers",
+    "stable_ids_unchanged_after_rollback": true,
+    "supervisor_restart": "last-known-good legacy stable release set preserved",
+    "candidate_recovery": "candidate cognitive-worker restored healthy",
+    "stable_ids_unchanged_after_candidate_recovery": true
+  },
+  "final_state": {
+    "traffic_mode": "STABLE",
+    "stable_upstream": "http://creator-web:8084",
+    "last_known_good_slot": "legacy-stable",
+    "public_health": 200,
+    "public_ready": 200
+  },
+  "limits": [
+    "The existing stable project remains legacy-stable during the non-disruptive transition; it was not recreated merely to rename it.",
+    "Local MEDIUM discovery was slow/unavailable during candidate cognitive startup; it delayed startup but did not affect stable traffic or the isolated release-set drill.",
+    "GEN2-COMPAT-OBS-001 has not resumed and no V1 path has been retired."
+  ]
+}