DEV Community

Discussion on: The importance of timeouts

orkon profile image
Alex Rudenko

A typical health check endpoint verifies that application integration points work correctly.

Why does it have to be so? My reasoning goes as follows: if the integration point is not working correctly, does it mean the service itself is unhealthy? No, because it can handle the integration point failure. Is there a need to shut down the service instance if an integration point is down? No, because a restart is unlikely to restore the integration point.

I have seen quite many times how services start to fail because of failing health checks due to, for example, a network connection that is 1 second slower than usual.