DEV Community

loading...

Discussion on: Let's talk about Health Checks

Collapse
phlash909 profile image
Phil Ashby

Nicely done! We are currently implementing health checks in our microservice platform at work, and have chosen to support two variants:

  • internal: similar to your description above, this ensures the service is able to accept requests itself.
  • external: as an extension to the internal check, this also passes on the healthcheck call to dependant services, ensuring the end-end chain of services are able to accept requests.

We added the external check for use by platform monitoring, avoiding the need to have internal network access, and as a way to dynamically record request routing, as we collect service IDs and versions in results. This helps with configuration drift detection / reconciliation in a complex service mesh platform.