Unmanage & unknown status impacting availability graph in SAM

I'm afraid that SolarWinds answer is correct. That is a way how built-in logic works. If you unmanaged your application then polling doesn't work so SAM cannot know if applications was UP or DOWN and it's status is UNKNOWN. I'd suggest 2 solutiions for your problem:

1. Use Unmute Alerts for the application, so you may see what actually happened with your application without alerting; unmuting alerts does not stop polling and SAM would show correct status on your chart.

2. Create custom chart (Custom HTML widget using SWQL) for application showing STATUS as NOT DOWN.

Hi JaroslawLadyga,

Thanks for your reply. But it will not solve the problem.

I know when my application was unmanaged. But question is solarwinds calculating the availability based on all status. It should be based on only Up and down status. Otherwise no use to unmanage the applications/nodes for planned activity if we it will not provide the 100% uptime report.

Hi alankar.srivastava,

I got your point but please consider how SAM would show you Up or DOWN status if it is not polling the application (or node - if node is UNMANGED then any polling doesn't work on it). SAM simply didn't know the status while application was UNMANAGED - that works the logic and it reflects on chart. If you want to show only Up or DOWN then you have to suppose that everything except of DOWN is UP which may be not true and looks not logical for me. Anyway if you want to build this chart you must use SWQL querry for that.

I am sure they could add the option somewhere to exclude "unknown" from the uptime calculation.  It can't hurt to submit that as a feature request.

The difficult part is that Unknown can come from other situations other than only unmanage.  So in reality unmanage really needs to set a status of unmanaged instead of unknown.  Then exclude unmanaged from the uptime calculation.

Ultimately though that would be faking uptime, true uptime comes from never unmanaging the service, only mute alerts during maintenance.  Accept the downtime and account for it.  If you are reporting 95% uptime (with 5% planned maintenance) and the execs what 99.999% uptime, then they need to commit the resources to have a clustered environment so that it's not down during maintenance.   Commit the resources or accept reality, everything else is a lie.

Yes, I understand that there are non-clusterable applications and situations that require complete downtime even for clustered environments, but that falls under "accept reality".

thanks for suggestion.

Submitted the Feature Request.