feat: add platform observability tracing CLI

2026-06-19 13:49:39 +00:00
parent a0dbf3d084
commit 15e85a1897
4 changed files with 1376 additions and 18 deletions
@@ -0,0 +1,122 @@
+version: 1
+kind: platform-infra-observability
+
+metadata:
+  id: observability
+  owner: unidesk
+  spec: PJ2026-01060501 OTel追踪 draft-2026-06-19-p0
+  relatedIssues:
+    - 489
+
+defaults:
+  targetId: D601
+
+images:
+  collector:
+    repository: otel/opentelemetry-collector-contrib
+    tag: 0.130.1
+    pullPolicy: IfNotPresent
+  tempo:
+    repository: grafana/tempo
+    tag: 2.8.1
+    pullPolicy: IfNotPresent
+
+targets:
+  - id: D601
+    route: D601:k3s
+    namespace: platform-infra
+    role: active
+    enabled: true
+    createNamespace: true
+
+collector:
+  deploymentName: otel-collector
+  serviceName: otel-collector
+  configMapName: otel-collector-config
+  replicas: 1
+  healthPort: 13133
+  otlp:
+    grpcPort: 4317
+    httpPort: 4318
+
+traceBackend:
+  type: tempo
+  deploymentName: tempo
+  serviceName: tempo
+  configMapName: tempo-config
+  replicas: 1
+  httpPort: 3200
+  otlp:
+    grpcPort: 4317
+    httpPort: 4318
+  storage:
+    mode: emptyDir
+    retention: 24h
+
+sampling:
+  mode: parentbased_traceidratio
+  ratio: 1
+
+instrumentation:
+  contextPropagation:
+    - tracecontext
+    - baggage
+  serviceConnections:
+    - serviceName: hwlab-cloud-api
+      owningRepo: pikasTech/HWLAB
+      targetNode: D601
+      lane: v0.3
+      namespace: hwlab-v03
+      requiredSpans:
+        - POST /v1/agent/chat
+        - durable_admission
+        - billing_preflight
+        - agentrun_dispatch
+        - projection_write
+        - trace_events_read
+        - turn_status_read
+    - serviceName: user-billing
+      owningRepo: pikasTech/HWLAB
+      targetNode: D601
+      lane: v0.3
+      namespace: hwlab-v03
+      requiredSpans:
+        - billing_preflight
+    - serviceName: agentrun-manager
+      owningRepo: pikasTech/agentrun
+      targetNode: D601
+      lane: v0.2
+      namespace: agentrun-v02
+      requiredSpans:
+        - agentrun_dispatch
+        - run_created
+        - command_result
+        - projection_sync
+
+resourceAttributes:
+  required:
+    - service.name
+    - deployment.environment
+    - unidesk.node
+    - hwlab.lane
+    - k8s.namespace.name
+    - git.commit
+  businessCorrelationAttributes:
+    - traceId
+    - sessionId
+    - turnId
+    - runId
+    - commandId
+
+probes:
+  readinessPath: /ready
+  traceQueryPathTemplate: /api/traces/{{traceId}}
+  statusEndpoints:
+    - name: tempo-ready
+      service: tempo
+      portName: http
+      path: /ready
+    - name: collector-health
+      service: otel-collector
+      portName: health
+      path: /