Knowledge baseSoftware

Monitoring is not the same as error tracking

4 September 20263 min read

Colloquially it is all called "monitoring": knowing whether your application works. But there are two different questions behind that one word, and mixing them up means missing exactly the outage that hurts. One asks whether the service is up; the other asks what went wrong while everything was up.

What monitoring answers

Monitoring looks at your system from the outside and from below: is the server reachable, is there disk space and memory, does the site respond within normal time, is the certificate still valid. The answer is a state, and the value is in the alarm: you hear it when something goes from healthy to unhealthy, preferably before users notice anything.

What error tracking answers

Error tracking looks from the inside: the application reports every error itself, with stack trace, release and context. That catches the outages monitoring is blind to — the payment button failing for everyone with an accent in their name, the export file that has been empty since the last release. The server is healthy, the site responds briskly, every light is green; yet something is broken for part of your users, and without error tracking you hear about it when someone calls.

One's blind spot is the other's territory

Monitoring sees the server die but not the exception; error tracking sees the exception but not the full disk. Stronger still: when the server goes down, error tracking goes down with it, because it runs on the application that just disappeared. These are not competitors to choose between, but two layers that cover each other's gaps. A healthy environment has both, and alerts from both should arrive in the same place, so nobody has to watch two dashboards.

Both, by default

Applications we build and run get both as part of delivery: monitoring on the infrastructure, error tracking in the application, alerts on one channel. If you honestly do not know what would be reported today if something in your application broke: get in touch and we will map it out together — or first read how we deliver software.

Frequently asked questions

Frequently asked questions

If my monitoring is green, does my application work?

Not necessarily. Monitoring looks at the service as a whole; an error that hits one feature or one group of users keeps every light green. For that you need error tracking, which reports from inside the code what went wrong.

Does error tracking replace my monitoring?

No. Error tracking only sees what the application itself can still report. A full disk, an expired certificate or a server that will not boot does not announce itself as a tidy exception; that is monitoring's territory.

What should I arrange first?

Whatever is missing now. Most environments have some monitoring and no error tracking, which means code errors are only found when a user calls. In that case error tracking is the quickest win.

Answer not found?

Ask an engineer directly — we usually respond within one business day.