[Vipergrün]

Agreed that SLO's and SLA's are needed. Typically these are tied to up-time and un-plannned downtime, such as you experienced.

Maintenance windows, or planned outages are critical to any enterprise environment. If upper management does not 'get it', you are in a bad situation. If your environment is setup correctly, you can have donwtime and have minimal impact on customers.

Only in rare cases should you not upgrade software/firmware. Just wait until they are hacked due to some security hole being exploited.

You should be able to enable debugging on your routers/switches/etc and look at that after the fact.....

Good luck..