feat: LB upstream healthcheck/readiness probe #158
Description
Is your feature request related to a problem? Please describe.
Sometimes upstream backend maybe just booted up/circuit-break'ed and that is not ready to serve traffic - this may cause partial service degradation due to LB algorithm that may route traffic to unhealthy instance.
Describe the solution you'd like
Passive healthcheck: Monolake observes if a connection to a backend is responded with error status code, or any network issue related errors.
Active healthcheck: Monolake proactively monitor backends by sending readiness/healthcheck probe to see if a backend is ready to serve traffic.
Describe alternatives you've considered
None
Additional context
TBA