Débogage CloudWatch et X-Ray AWS

---

name: capability:cloud-watch:debug description: Analyze AWS CloudWatch and X-Ray telemetry — surface errors, latency issues, trace anomalies, and alarm violations argument-hint: "--service-name <name> [--time-range <hours>] [--description <text>]" user-invocable: true agent: zzaia-devops-specialist metadata: parameters: - name: service-name description: AWS service name or log group to query required: true - name: time-range description: Time range in hours to analyze (default 24) required: false default: "24" - name: description description: Additional context or instructions required: false

PURPOSE

Retrieve and analyze AWS CloudWatch and X-Ray telemetry for a service. Surfaces errors, latency issues, trace anomalies, alarm violations, and quality problems. Returns structured diagnostic findings — not raw data.

EXECUTION

1. Retrieve Telemetry — Call /capability:cloud-watch:query --service-name <service-name> --time-range <time-range>

   • CloudWatch logs and metrics
   • CloudWatch alarms and states
   • X-Ray traces and Application Signals
   • Trace segments and service nodes

2. Analyze — Runtime Issues

   • Group errors by type and message; calculate frequency
   • Identify failed traces and root cause services
   • Flag alarms in ALARM or INSUFFICIENT_DATA states
   • Map occurrences to timeline

3. Analyze — Quality Issues

   • Detect error rate exceeding acceptable threshold (> 1% of requests)
   • Flag SLA violations: trace latency above defined targets
   • Identify throughput degradation trends across the time range
   • Surface slow spans in traces and bottleneck services
   • Flag alarm violations and metric threshold breaches

4. Return Findings — Structured diagnostic output with severity-tagged issues and quality violations

DELEGATION

MANDATORY: Always invoke the agents defined in this command's frontmatter for their designated responsibilities. Never skip, replace, or simulate their behavior directly.

• zzaia-devops-specialist — Query CloudWatch and X-Ray MCPs and analyze diagnostic data

WORKFLOW

sequenceDiagram
    participant U as User
    participant C as Command
    participant A as DevOps Agent
    participant CW as CloudWatch MCP
    participant XR as X-Ray MCP

    U->>C: /capability:cloud-watch:debug --service-name <service>
    C->>A: Dispatch with service filter and time range
    A->>CW: Query logs, metrics, alarms
    A->>XR: Query traces and Application Signals
    CW-->>A: CloudWatch data
    XR-->>A: X-Ray data
    A->>A: Group errors, analyze traces, identify slow spans, flag alarms
    A-->>C: Structured diagnostic findings
    C-->>U: Issues, latency problems, anomalies, alarm violations

ACCEPTANCE CRITERIA

• Telemetry retrieved for the specified time range from both MCPs
• Errors deduplicated and grouped by type with frequency
• Failed traces identified with root cause services
• Alarms in ALARM state flagged with severity
• Slow spans and latency issues identified in traces
• Quality violations identified: error rate, SLA breaches, throughput degradation, alarm violations
• Timestamps included for all critical events

EXAMPLES

/capability:cloud-watch:debug --service-name payment-service

/capability:cloud-watch:debug --service-name api-gateway --time-range 48

/capability:cloud-watch:debug --service-name lambda-processor --time-range 12 --description "Check for timeout issues"

OUTPUT

• Issues: Error types, frequency, affected services
• Traces: Failed traces with root cause analysis
• Latency: Slow spans, bottleneck services, SLA violations
• Alarms: Active alarms with threshold and state
• Anomalies: Detected trace anomalies and deviations
• Performance: Request latency, throughput, error rate trends
• Quality Violations: Error rate threshold breaches, SLA violations, throughput degradation, alarm violations
• Timeline: Chronological summary of key events