What is considered a failure to trigger a retry?

[creasewp]

I have been unable to find any documentation that indicates what exactly will be considered a failure that will trigger a retry. I assume things like a timeout or pod being unavailable would be considered a failure, but what about HTTP status codes? Through testing, I figured out that HTTP 500 is a failure, but what about others? Are all 5xx errors considered failures? What about 4xx?

[mateiidavid]

Hey @creasewp, by default only 5xx status codes are considered failures; a 403 for example (Forbidden) would be classed as a success. With that being said, you can overwrite the defaults on a service-by-service basis using ServiceProfiles and adding in a responseClass object.