architect

To find observability gaps in a Cloudflare Workers codebase, you audit recent changes to validate that logs, traces, and error signals are actionable, ensuring structured evlog-style logging is used to prevent blind diagnosis.

The best way to track codebase health issues is to use a continuous task-system hygiene approach, using the repository `tasks/` folder to locate existing work, append new evidence to matching tasks, and create new task files only when needed.

To diagnose performance regressions in Cloudflare Worker services, you apply continuous monitoring workflows using MCP-based observability tools like Workers Observability and PostHog to detect anomalies and guide targeted remediation.

Yes, this approach works for continuous monitoring of repository-wide changes by proactively diagnosing and remediating observability gaps, error regressions, flaky tests, and performance issues before they become chronic production incidents.

You need structured evlog-style logging for error diagnosis because it ensures logs, traces, and error signals are actionable, preventing blind diagnosis when investigating error regressions and worker debugging issues.