topic Re: How to make a containers recycle looks clean? in Container platforms

How to make a containers recycle looks clean?

waikeat_chan — Mon, 13 Dec 2021 13:18:55 GMT

For the containerized service, everything looks fine (looks fine in the sense that, every time the underlying docker is recycle, I didn't see any gap or broken pieces in the metric graph in service layer)

However, in host layer. The old containers would always goes into the state of 'unmonitored'.

For example, as this screenshot shown, api-private-gateway would always get its name increment by one each time it recycle. Right now, api-private-gateway-64-* is running, is there anyway I can make the containers before 64 disappear? (or, if they are gonna still be there can their state change from 'unmonitored' to something more......meaningful per se?)

Best Regards,

Wai Keat

Re: How to make a containers recycle looks clean?

ChadTurner — Thu, 16 Apr 2020 14:38:27 GMT

There is a new method where you need to pass an event within 60mins of the recycle which will then keep you from being alerted when a host is being spun down in cloud environments. You can read more about it here: https://www.dynatrace.com/news/blog/new-event-type-helps-avoid-unnecessary-alerts-for-planned-host-downscaling/

Re: How to make a containers recycle looks clean?

Enrico_F — Thu, 16 Apr 2020 15:06:20 GMT

That's neat but doesn't really help addressing the "unmonitored" PaaS hosts spam for terminated containers/pods in an application-only monitoring scenario (= oneagent is manually deployed into the container).

Re: How to make a containers recycle looks clean?

ChadTurner — Thu, 16 Apr 2020 15:14:20 GMT

Correct, and until Dynatrace has a solution for these scenarios.