Every item below is live in production right now. Mint a token, click around, prove it works.
✓Multi-region monitoring
A check from one region briefly failing was opening false incidents. Now an incident only opens when consensus says down. Five regions: Western Europe, Eastern Europe, Eastern North America, Asia Pacific, Oceania.
✓Public platform status page
We monitor our own API, dashboard, and marketing site with 24Observe and publish the page so you can see if our service is up before you sign in. status.24observe.com.
✓Auto-extracted log facets
Every primitive JSON field in your log attrs becomes a click-to-filter chip in the sidebar. No schema to declare. The thing that turns "we have logs" into "we have observability".
✓Pattern grouping
Click "Patterns" on the Logs page. 10,000 lines that differ only by a timestamp and a request ID collapse to one row with the template, the sample, and the count. The single biggest reason teams stay on Datadog — now here.
✓Error tracking
Recurring stack traces auto-group by signature. JavaScript, Python, Java, Go, generic ERROR/FATAL. Same logical error across different deploys, paths, line numbers — same row. Bundles the thing Sentry charges a separate seat for.
✓Log-derived metrics
Save any log search as a time-series. Chartable, alertable, auto-refreshing every 60 seconds. Datadog charges 10¢ per metric per month for this. Here it is free.
✓Anomaly alerts on log volume
Threshold alerts when you know the magic number. Anomaly alerts when "normal" varies. Compares current window vs 7-day rolling baseline, fires at 3× spike. No ML configuration, no model tuning.
✓CloudWatch + Heroku + Vercel + Docker + journald
Four new "no SDK to install" sources land in one paste. Lambda logs flow through Kinesis Firehose. Heroku in one drains:add. Vercel in one URL. Docker containers via the syslog driver. systemd via journald. Twelve sources total now.