observability

Observability

Part of Agent Skills™ by googleadsagent.ai™

Description

Observability implements structured logging, distributed tracing, and error monitoring for Cloudflare Workers and edge applications. The agent instruments code with contextual log entries, trace spans, error boundaries, and alerting rules that provide full visibility into production behavior without sacrificing performance.

Workers present unique observability challenges. There are no persistent processes to attach profilers to, no filesystem for log files, and no APM agent to inject. Observability must be built into the application layer through structured log events shipped to external collectors, request-scoped trace contexts, and explicit error capture with stack traces and request metadata.

This skill covers three pillars: Logging (structured JSON events with correlation IDs), Tracing (request-scoped spans with timing and metadata), and Monitoring (error rate tracking, latency percentiles, and alerting thresholds). Together, they answer the three questions of production debugging: what happened, how long did it take, and how often does it fail.

Use When

• Adding logging to Workers or edge applications
• Debugging production issues with distributed request tracing
• Setting up error monitoring and alerting
• Implementing health check endpoints
• Measuring and reporting latency percentiles
• Building dashboards for operational visibility

How It Works

graph TD
    A[Incoming Request] --> B[Generate Trace ID]
    B --> C[Start Root Span]
    C --> D[Execute Handler]
    D --> E[Child Spans for Subrequests]
    E --> F{Error?}
    F -->|Yes| G[Capture Error + Context]
    F -->|No| H[Record Success Metrics]
    G --> I[Structured Log: ERROR]
    H --> J[Structured Log: INFO]
    I --> K[Ship to Collector via waitUntil]
    J --> K
    K --> L[External: Datadog / Grafana / Sentry]
    C --> M[End Root Span + Record Latency]
    M --> K