Accurate Availability - Using Retrospective Scheduled Maintenance

markgosling

Looking for the communities input/methods on accurate Availability reporting.

How do you retrospectively accommodate Planned Outages when the Solarwinds monitoring of the elements has not been placed into Scheduled Maintenance Mode?

The scenario is that some building works or the like has occurred, impacted the availability of nodes etc but these nodes where still under active Solarwinds monitoring. Naturally Solarwinds correctly reports these as down and decrements the Availability for those Nodes. Problem is the business 'knows' these are not real outages as the building works/etc were planned hence the rolled up Availability figures will now be inaccurate.

What do you do?

• Exclude the affected nodes from the monthly rolled up Availability figures?
• Schedule the nodes into Maintenance after the fact? - Don't think Solarwinds recalculates the Nodes availability though
• Manually adjust the Availability figures (how?)
• Use the Planned Works to explain the reduced availability number?
• Do something else that is as cunning as a rat with a gold tooth?

In a perfect world every piece of disruptive work would be scheduled, communicated and excluded from monitoring via Scheduled Maintenance feature. Of course our world is less than perfect.

All input/suggestions gratefully accepted.

tonymckay

We've run into the same issue where an agent has forgotten a network element that would be impacted by scheduled maintenance.  We've found that putting the node into maintenance mode/unmanage once the node is down, then the availability will still show 0% for the duration, having no beneficial effect on availability numbers.  We either catch it before it goes down or scrub the metric out of the aggregate.  It's not pure, but it's repeatable.  I think a retrograde polling period exclusion would be a massive game-changer in this aspect.