Cloud & Infrastructure
Sentry Development
Knowing an error happened before the customer tells you.
The gap between an error happening and someone knowing about it is where reputation is lost. Logs record everything and are read by nobody; Sentry inverts that — it groups errors, tells you how many users are affected, and carries the stack trace, request and release that produced them.
That context is the difference between an alert and a fix. Knowing an exception occurred is not useful; knowing it started with yesterday's release and affects a specific action is.
Why we use it
Sentry is in every production application we operate. It is the first thing configured after a deployment pipeline, because an application without error visibility is one you find out about from a customer.
We tune it deliberately: alerts on new and spiking issues rather than on every occurrence. An alerting channel that fires constantly gets muted, and a muted channel is worse than none because it feels like coverage.
Planning a Sentry project?
We will tell you honestly whether it is the right choice for what you are building.