Every failure becomes an issue, not another log line
A busy API can log thousands of errors an hour that come down to three causes. Midline groups occurrences of the same failure onto one issue, tells you when it appears, and reopens it if you resolve it and it returns.
What is grouped into issues
Grouping runs on ingest, so it covers whatever reaches Midline from the server agent, the browser SDK or a plain HTTP post.
- Error events and security events.
- Any request answered with a 5xx, even when your framework caught the exception itself. This matters on NestJS, where exceptions are handled by filters and rarely reach Express error middleware.
How two errors become one issue
Each event is fingerprinted from its event type, method, route and message, with ids and numbers folded out. "User 42 not found" and "User 91 not found" are the same problem, so they land on the same issue and the count goes up instead of the list growing.
A lifecycle that matches how teams work
An issue is unresolved, resolved or ignored, with a record of who changed it and when. Resolve it after a fix and, if the same failure fires again, it reopens instead of staying buried.
Who hears about it
Organisation owners and admins get an email when a new issue appears and when a resolved one comes back. Starter includes 4 issue-alert emails a day; Basic and Pro are unlimited.
One place for the API and the page that called it
The browser SDK reports uncaught errors, unhandled rejections and failed fetch calls from the page. Same-origin calls carry a traceparent header, so a failed call in the browser and the request behind it share a trace id.
Questions
What is the difference between error tracking and logging?
A log is a stream of individual lines. Error tracking collapses repeats into a single issue with a count, a status and a history, so you see three problems instead of three thousand lines.
Does Midline track 5xx responses as well as thrown exceptions?
Yes. Any request answered with a 5xx is grouped into an issue, even if the framework handled the exception and no error ever reached your own error handler.
What happens when a resolved issue happens again?
It reopens, and Midline emails organisation owners and admins so a regression does not go unnoticed.
Does Midline replace a full observability platform?
No, and it is not trying to. Midline is narrower by design: requests, errors and console output for an API and the web app that calls it, grouped into issues. It suits teams that want that without operating a larger stack.
Quick start · Browser SDK · Node.js agent
Start free