Health check with exponential back off on startup
Closed this issue · 1 comments
jasonjoo2010 commented
Use case
Assume the health check interval to T.
Assume the initial health check failed once receiving the started event from docker.
In the current implementation, the second chance to report its health will be after T time.
While in some large IDCs we have a significant long interval (T) and it could be more than 100 seconds.
Suggestion
When health check booted it may make more sense to proceed with a backoff interval up to T with initial unhealth.