AWS Load balancer health checks

[opethdisciple]

Any one have much experience with AWS load balancers and health checks?

I am working on a NLB at work using TCP and we are not sure if it is a good idea to put our health checks to the fastest they go i.e 2 health checks with a 10 second interval.

The default health check is 3 intervals of 30 seconds.

The reason for the fear is we are thinking maybe this might also cause false positives and mark the node unhealthy if the TCP port check has any sort of brief issues.

To give more information we have two databases behind the loadbalancer in two different azs.

The idea is not to load balance but to improve availability.

 

[ubern00b]

It all depends on what you deem as acceptable downtime. 2 checks 10 second interval means a possible 20sec of downtime before it will mark the endpoint as down. 3 x 30, you get the idea. There is no "good idea" or "bad idea". It is about what works for your implementation.

Saying that, if you are worried about generating "false positives" the only way that happens is if the health check gets something other than a 200 in the allotted time and if it is then what are your end users experiencing?