apiVersion: platform.chatcli.io/v1alpha1
kind: Instance
metadata:
  name: chatcli-prod
spec:
  provider: CLAUDEAI
  replicas: 1
  apiKeys:
    name: chatcli-api-keys
  server:
    port: 50051
    tls:
      enabled: true
      secretName: chatcli-tls       # Secret with tls.crt, tls.key, ca.crt (optional)
    token:
      name: chatcli-server-token    # Secret containing the auth token
      key: token                    # Key within the Secret (default: "token")
    security:
      rateLimitRps: 20
      rateLimitBurst: 50
      bindAddress: "0.0.0.0"
      jwtSecretRef:
        name: chatcli-jwt
        key: secret
  watcher:
    enabled: true
    interval: "30s"
    targets:
      - name: api-gateway
        namespace: production
        metricsPort: 9090
      - name: worker
        namespace: batch
      - name: postgres                # StatefulSet monitoring
        kind: StatefulSet
        namespace: production
      - name: fluentd                 # DaemonSet monitoring
        kind: DaemonSet
        namespace: logging
      - name: etl-pipeline            # CronJob monitoring
        kind: CronJob
        namespace: data
  agents:
    configMapRef: chatcli-agents         # ConfigMap with agent .md files
    skillsConfigMapRef: chatcli-skills   # ConfigMap with skill .md files
  plugins:
    image: myregistry/chatcli-plugins:latest  # Init container with plugin binaries
    # pvcName: chatcli-plugins-pvc           # Or use an existing PVC
  extraEnv:
    - name: CHATCLI_AGENT_SECURITY_MODE
      value: "strict"

apiVersion: platform.chatcli.io/v1alpha1
kind: Instance
metadata:
  name: chatcli-copilot
spec:
  provider: COPILOT
  model: gpt-4o          # or gpt-4o-mini, claude-sonnet-4, gemini-2.0-flash
  replicas: 1
  apiKeys:
    name: chatcli-copilot-keys   # Secret with GITHUB_COPILOT_TOKEN
  server:
    port: 50051

apiVersion: v1
kind: Secret
metadata:
  name: chatcli-copilot-keys
type: Opaque
stringData:
  GITHUB_COPILOT_TOKEN: "ghu_xxxxxxxxxxxx"   # From /auth login github-copilot
  # COPILOT_MODEL: "gpt-4o"                   # Optional model override
  # COPILOT_API_BASE_URL: "https://..."        # Optional enterprise URL

apiVersion: platform.chatcli.io/v1alpha1
kind: Instance
metadata:
  name: chatcli-bedrock
spec:
  provider: BEDROCK
  # Modern Claude on Bedrock requires an inference profile id
  # (prefix global./us./eu./apac.). For Claude 3/3.5 the base id also works.
  model: global.anthropic.claude-sonnet-4-5-20250929-v1:0
  replicas: 1
  apiKeys:
    name: chatcli-bedrock-keys
  server:
    port: 50051

apiVersion: v1
kind: Secret
metadata:
  name: chatcli-bedrock-keys
type: Opaque
stringData:
  BEDROCK_REGION: "us-east-1"

apiVersion: v1
kind: Secret
metadata:
  name: chatcli-bedrock-keys
type: Opaque
stringData:
  AWS_ACCESS_KEY_ID: "AKIA..."
  AWS_SECRET_ACCESS_KEY: "..."
  # AWS_SESSION_TOKEN: "..."                   # optional, for STS
  BEDROCK_REGION: "us-east-1"
  # Corporate proxy / private CA (optional):
  # CHATCLI_BEDROCK_CA_BUNDLE: "/etc/ssl/corp-ca-bundle.pem"

apiVersion: platform.chatcli.io/v1alpha1
kind: Anomaly
metadata:
  name: watcher-highrestartcount-api-gateway-1234567890
  namespace: production
spec:
  signalType: pod_restart      # pod_restart, oom_kill, pod_not_ready, deploy_failing
  source: watcher
  severity: warning
  resource:
    kind: Deployment
    name: api-gateway
    namespace: production
  description: "HighRestartCount on api-gateway: container app restarted 8 times"
  detectedAt: "2026-02-16T10:30:00Z"

apiVersion: platform.chatcli.io/v1alpha1
kind: Issue
metadata:
  name: api-gateway-pod-restart-1771276354
  namespace: production
spec:
  severity: high
  source: watcher
  signalType: pod_restart       # Propagated from Anomaly for tiered Runbook matching
  description: "Correlated incident: pod_restart on api-gateway"
  resource:
    kind: Deployment
    name: api-gateway
    namespace: production
  riskScore: 65
status:
  state: Analyzing
  remediationAttempts: 0
  maxRemediationAttempts: 3

apiVersion: platform.chatcli.io/v1alpha1
kind: AIInsight
metadata:
  name: api-gateway-pod-restart-1771276354-insight
  namespace: production
spec:
  issueRef:
    name: api-gateway-pod-restart-1771276354
  provider: CLAUDEAI
  model: claude-sonnet-4-5
status:
  analysis: "High restart count caused by OOMKilled. Container memory limit (512Mi) insufficient for current workload."
  confidence: 0.87
  recommendations:
    - "Increase memory limit to 1Gi"
    - "Investigate memory leak in application"
  suggestedActions:
    - name: "Restart deployment"
      action: RestartDeployment
      description: "Restart pods to reclaim leaked memory"
    - name: "Scale up replicas"
      action: ScaleDeployment
      description: "Add replicas to distribute memory pressure"
      params:
        replicas: "4"
  generatedAt: "2026-02-16T10:31:00Z"

apiVersion: platform.chatcli.io/v1alpha1
kind: RemediationPlan
metadata:
  name: api-gateway-pod-restart-1771276354-plan-1
  namespace: production
spec:
  issueRef:
    name: api-gateway-pod-restart-1771276354
  attempt: 1
  strategy: "Attempt 1 via runbook 'auto-pod-restart-high-deployment'. AI analysis: High restart count caused by OOMKilled..."
  actions:
    - type: RestartDeployment
    - type: ScaleDeployment
      params:
        replicas: "4"
  safetyConstraints:
    - "No delete operations"
    - "No destructive changes"
    - "Rollback on failure"
status:
  state: Completed
  result: "Deployment restarted and scaled to 4 replicas successfully"

apiVersion: platform.chatcli.io/v1alpha1
kind: Runbook
metadata:
  name: high-error-rate-runbook
  namespace: production
spec:
  description: "Standard procedure for high error rate on Deployments"
  trigger:
    signalType: error_rate
    severity: high
    resourceKind: Deployment
  steps:
    - name: Scale up deployment
      action: ScaleDeployment
      description: "Increase replicas to handle error spike"
      params:
        replicas: "4"
    - name: Rollback if needed
      action: RollbackDeployment
      description: "Rollback to previous version if errors persist"
  maxAttempts: 3

apiVersion: platform.chatcli.io/v1alpha1
kind: RemediationPlan
metadata:
  name: api-gateway-pod-restart-plan-1
  namespace: production
spec:
  issueRef:
    name: api-gateway-pod-restart-1771276354
  attempt: 1
  strategy: "Agentic AI remediation"
  agenticMode: true
  agenticMaxSteps: 10
  agenticHistory:
    - stepNumber: 1
      aiMessage: "High restart count with OOMKilled. Scaling up to reduce memory pressure."
      action:
        type: ScaleDeployment
        params:
          replicas: "5"
      observation: "SUCCESS: ScaleDeployment executed successfully"
      timestamp: "2026-02-16T10:31:00Z"
    - stepNumber: 2
      aiMessage: "Pods still restarting. Adjusting memory limits to 1Gi."
      action:
        type: AdjustResources
        params:
          memory_limit: "1Gi"
          memory_request: "512Mi"
      observation: "SUCCESS: AdjustResources executed successfully"
      timestamp: "2026-02-16T10:31:35Z"
    - stepNumber: 3
      aiMessage: "All pods running stable. Issue resolved."
      timestamp: "2026-02-16T10:32:10Z"
status:
  state: Completed
  agenticStepCount: 3

apiVersion: platform.chatcli.io/v1alpha1
kind: PostMortem
metadata:
  name: pm-api-gateway-pod-restart-1771276354
  namespace: production
spec:
  issueRef:
    name: api-gateway-pod-restart-1771276354
  resource:
    kind: Deployment
    name: api-gateway
    namespace: production
  severity: high
status:
  state: Open
  summary: "OOMKilled containers caused cascading restarts on api-gateway"
  rootCause: "Memory limit (512Mi) insufficient for current workload pattern"
  impact: "Service degradation for 5 minutes, 30% error rate increase"
  timeline:
    - timestamp: "2026-02-16T10:30:00Z"
      type: detected
      detail: "Issue detected: pod_restart on api-gateway"
    - timestamp: "2026-02-16T10:31:00Z"
      type: action_executed
      detail: "ScaleDeployment to 5 replicas"
    - timestamp: "2026-02-16T10:31:35Z"
      type: action_executed
      detail: "AdjustResources memory_limit=1Gi"
    - timestamp: "2026-02-16T10:32:10Z"
      type: resolved
      detail: "All pods stable, issue resolved"
  actionsExecuted:
    - action: ScaleDeployment
      params:
        replicas: "5"
      result: success
      timestamp: "2026-02-16T10:31:00Z"
    - action: AdjustResources
      params:
        memory_limit: "1Gi"
        memory_request: "512Mi"
      result: success
      timestamp: "2026-02-16T10:31:35Z"
  lessonsLearned:
    - "Memory limits should account for peak workload patterns"
    - "Set up HPA to auto-scale on memory pressure"
  preventionActions:
    - "Configure HPA with min 3 replicas for api-gateway"
    - "Set memory limit to 1Gi across all environments"
  duration: "2m10s"
  generatedAt: "2026-02-16T10:32:10Z"