What problem does it solve? When the Cloudflare-to-Datadog analytics sync cron goes down for more than 30 minutes, it leaves a permanent gap in the openrouter.cloudflare.* metrics because the cron only ever queries the last 30 minutes. This Skill walks you through finding the gap, replaying the missing range, and verifying the points actually landed. ## Core Features & Use Cases - Gap Detection: Query Datadog metrics to locate the exact boundaries of the missing data and confirm the sync has recovered before replaying. - Guided Backfill: Run scripts/cf-analytics-backfill.ts with Infisical-managed credentials over half-open, whole-minute ranges split into 30-minute chunks, with automatic retries and per-window failure reporting. - Verification & Latency Awareness: Understand Datadog historical metrics ingestion (HMI) requirements and latency windows so you don't re-run unnecessarily. - Use Case: The [CF Analytics] GraphQL to Datadog sync is dead monitor fires after a two-hour outage; use this Skill to identify the gap, backfill it safely, and confirm the graph is whole again. ## Quick Start Ask the agent to backfill the Cloudflare analytics gap in Datadog for the affected time range after the cf-analytics-sync outage.
