feat(aiops): add typed CPU P99 recovery closure
This commit is contained in:
@@ -1,7 +1,7 @@
|
||||
{
|
||||
"schema_version": "sre_ai_agent_conversation_commitments_v1",
|
||||
"program_id": "AIA-SRE-P0-20260715",
|
||||
"generated_at": "2026-07-22T12:33:42+08:00",
|
||||
"generated_at": "2026-07-22T15:36:56+08:00",
|
||||
"source_scope": "This Codex AI Agent task only; cross-task delegation envelopes are excluded.",
|
||||
"authoritative_for_execution_order": false,
|
||||
"raw_conversation_embedded": false,
|
||||
@@ -51,7 +51,7 @@
|
||||
|
||||
{"id":"AIA-CONV-032","category":"named_incident","status":"source_implemented_runtime_pending","title":"INC-20260711-11C751 cold-start-gate 補 PlayBook、transport、rollback 與 verifier","linked_work_items":["AIA-SRE-004","AIA-SRE-006","AIA-SRE-015"],"terminal_condition":"namespace identity 修正後完成一次 bounded same-run repair 或明確 no-write terminal。","source_evidence":["source namespace separated from normalized host_recovery execution identity","exact inventory scope absence blocks Recover before single-flight claim or Agent99 transport","durable alert operation receipt records the explicit no-write terminal","consumer-validated source no-write verifier receipt with runtime closure false"]},
|
||||
{"id":"AIA-CONV-033","category":"named_incident","status":"source_implemented_runtime_pending","title":"INC-20260711-D037E5 修復 stale candidate 阻擋、自動排隊與 durable closure","linked_work_items":["AIA-SRE-007","AIA-SRE-015"],"terminal_condition":"新候選可 claim,apply 不重複,verifier、Telegram 與 learning receipts 同 run 完成。","source_evidence":["D037E5 canonical workload identity resolves to the Kubernetes executor/verifier without consulting legacy Ansible candidates","fingerprint-stable durable candidate event supports pending and pre-dispatch lease-expired atomic claim","dedicated feature-gated Kubernetes executor job owns a bounded one-candidate claim tick and is not started by API, signal worker or Ansible broker","single-dispatch barrier suppresses duplicate apply after dispatch starts","durable closure requires lifecycle, independent verifier, Telegram provider acknowledgement and learning receipts on one execution run"]},
|
||||
{"id":"AIA-CONV-034","category":"named_incident","status":"in_progress","title":"AWOOOI CPU 高負載與 P99 上升完成 RCA、bounded repair 與 verifier","linked_work_items":["AIA-SRE-004","AIA-SRE-006","AIA-SRE-010"],"terminal_condition":"相關 metrics/logs/changes 對齊 canonical asset,修復後 latency 與資源 verifier 關閉。"},
|
||||
{"id":"AIA-CONV-034","category":"named_incident","status":"source_implemented_runtime_pending","title":"AWOOOI CPU 高負載與 P99 上升完成 RCA、bounded repair 與 verifier","linked_work_items":["AIA-SRE-004","AIA-SRE-006","AIA-SRE-010"],"terminal_condition":"相關 metrics/logs/changes 對齊 canonical asset,修復後 latency 與資源 verifier 關閉。","source_evidence":["same-run canonical CPU/P99 metrics, sanitized log and recent-change receipt correlation","exact typed Kubernetes, Host Ansible and database RCA lanes with unknown-asset fail-closed behavior","fingerprint-deduplicated durable candidate receipt without provider, Agent99 or runtime mutation","durable closure requires typed execution plus deterministic independent resource and SigNoz P99 verifier receipts on one run"]},
|
||||
{"id":"AIA-CONV-035","category":"named_incident","status":"in_progress","title":"Backup/restore escrow 補 freshness、escrow metadata、restore drill 與 DR scorecard","linked_work_items":["AIA-SRE-011","AIA-SRE-016"],"terminal_condition":"read-only evidence 齊全;缺 escrow/restore receipt 時保持 blocked 而非假綠。"},
|
||||
{"id":"AIA-CONV-036","category":"named_incident","status":"source_implemented_runtime_pending","title":"Sentry Snuba DockerContainerUnhealthy 反覆告警自動修復","linked_work_items":["AIA-SRE-005"],"terminal_condition":"exact-container replay 完成 bounded recovery、健康 verifier 與 recurrence fence。"},
|
||||
{"id":"AIA-CONV-037","category":"named_incident","status":"source_implemented_runtime_pending","title":"AlertChainBroken_Alertmanager webhook error 大於 10% 自動旁路與修復","linked_work_items":["AIA-SRE-014","AIA-SRE-017"],"terminal_condition":"integration-scoped counters、旁路 delivery、receiver contract 與恢復 receipt 全部成立。"},
|
||||
@@ -98,8 +98,8 @@
|
||||
"active_or_completed_commitments": 68,
|
||||
"by_status": {
|
||||
"analysis_or_governance_complete": 6,
|
||||
"source_implemented_runtime_pending": 31,
|
||||
"in_progress": 29,
|
||||
"source_implemented_runtime_pending": 32,
|
||||
"in_progress": 28,
|
||||
"not_started_or_no_current_evidence": 2,
|
||||
"superseded": 2
|
||||
},
|
||||
|
||||
Reference in New Issue
Block a user