TL;DR: Put the curl command in the runbook so on-call does not have to remember it. ## 3. # Workflow: observability setup ## 1. Logs for debugging -

## Fix
1. Put the curl command in the runbook so on-call does not have to remember it. ## 3.
   Expected: this specific failure stops.

## Details
# Workflow: observability setup ## 1. Logs for debugging - `wrangler tail` during development and deploys: live console and exception logs, including Durable Object requests. - Real-time logs in the dashboard for when a terminal is not available. - Tail Workers (beta) for a persistent pipeline: filter, sample, and forward telemetry to your own systems. ## 2. The incident runbook Document the `cf-error-type` WAF visibility Review Security events regularly: which rules fire, what they catch, what they wrongly block. Every false positive is a tuning task; every gap is a new rule. ## 4. Cache visibility in deploys The deploy pipeline checks `CF-Cache-Status` on key URLs: MISS then HIT after invalidation. A deploy that nobody verified at the edge is a deploy that might not have taken effect for users. ## 5. Origin-side logging Restore the real client IP at the origin ( `CF-Connecting-IP` ) so origin logs are meaningful. Correlate origin logs with edge events by timestamp during incidents. ## Checklist - Tail during every deploy; runbook with the header triage; WAF events reviewed weekly. - Origin logs carry the real client IP, not Cloudflare's. - Log retention covers at least one full incident-review cycle.

## When to use
You hit exactly this: Workflow: observability setup for a Cloudflare zone and Workers in Workflow.

## When not to use
A different error, or the same symptom in a different tool. This page only covers the failure above.

## Compatibility
Workflow.

## Root cause
triage: headers absent means the origin errored; 1101/1102 means the Worker; 52x means origin connectivity; 525/526 means origin TLS.