Supabase Log Drains to Datadog: Setup, Parsing & Alerting
Supabase log drains ship your auth, Postgres, and edge function logs to Datadog in real time. This guide walks through setup, timestamp parsing, and the alerts that actually matter.
<!-- BRIEF SPEC (supabase-log-drains-observability) Variant: DEEP-GUIDE | Hook: CONTRARIAN | FAQ: 8 | ~1600-2300 words Cover style: DARK-NEON (deep-guide) — dark navy, neon cyan+magenta; abstract 3D database cylinder emitting a glowing log stream into a funnel. Distinct shape from the waitUntil cover. No text. In-body images (2): (1) architecture diagram: Supabase project -> batched gzip log drain -> Datadog intake -> monitors/alerts; (2) screenshot-style mock of a Datadog log stream showing parsed Supabase auth + postgres events with timestamps. SERP Step 0: official docs + one small-site setup guide (adwaitx) + skill READMEs — no full tutorial with parsing + alerting. Gap: real CLI walkthrough, Grok timestamp parsing, concrete auth/slow-query monitors, decision table vs analytics buckets vs metrics API. --> Last updated: 2026-10-10 Your Supabase dashboard logs are not observability. They are a demo of observability: one day of retention on the free plan, no real alerting, no correlation with the rest of your stack, and a search box that forgets everything yesterday. The moment a user reports "login was broken at 3am" and you cannot answer what happened, the dashboard has failed you. Supabase log drains fix this at the source — they stream your auth, Postgres, Realtime, and edge function logs to your own observability platform in near real time, where you can alert, graph, and retain them like a grown-up. This guide sets up a Datadog drain end to end, then builds the two alert families that catch real production incidents. What you will build and alert on A Supabase log drain shipping to Datadog, configured via the CLI ( supabase log-drains add/list/remove ), with correct timestamp parsing so events sort and graph properly. An auth-anomaly monitor that pages you when…