observability

Instrument code with structured logs, health endpoints, and persisted failure state.

Updated May 25, 2026
One-click install
npx skills add https://github.com/cmetech/otto-cli --skill observability-cmetech
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: observability
Source: https://github.com/cmetech/otto-cli/tree/main/src/resources/skills/observability
Command: npx skills add https://github.com/cmetech/otto-cli --skill observability-cmetech

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Observability is often neglected, leaving production issues undiagnosed until it's too late. This skill provides a blueprint to instrument code with signals, health endpoints, and durable failure state so incidents are diagnosable at 3am.

Core Features & Use Cases

  • Structured decision-point logs: log the what, why, and flowId at critical decision points rather than generic "entered function" messages.
  • Persistent failure state: write failure context to disk so a fresh agent can continue debugging without rerun.
  • Health and status surfaces: expose lightweight health checks or status files for long-running processes.

Quick Start

Use this skill to instrument a code path by adding structured logs, health endpoints, persisted failure state, and explicit failure modes.

Frequently Asked Questions about observability

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I add structured logging to long-running services for debugging at 3am?

Structured logging for long-running services requires logging the what, why, and flowId at critical decision points, providing actionable signals during production failures instead of generic function entry messages.

What is persistent failure state and how does it help debug background tasks?

Persistent failure state is writing failure context to disk so a fresh agent can continue debugging without rerun, ensuring background task incidents remain diagnosable after a process restarts.

Can I use health endpoints to monitor production failure states in auto-mode code?

Yes, you can expose lightweight health checks or status files for long-running processes, enabling production monitoring of explicit failure modes and durable state persistence in auto-mode style code.

What's the best way to make production failures visible in real time?

Making failures visible in real time is best achieved by instrumenting code with structured decision-point logs, durable state persistence, lightweight health surfaces, and explicit failure modes.

Why does my production monitoring fail to diagnose background task issues without code reruns?

Production monitoring fails to diagnose issues without reruns when code lacks durable failure state persistence, meaning critical failure context is lost when the process stops and cannot be resumed.