I think you will find that it kicks over as a result of the poller not being able to perform a database write during the maintenance (the hot-standby looks at the last database write timestamp).
Two things you could do...
Work out what the database issue is. Our pollers continue to perform database writes during maintenance runs - indicates to me that something is holding an exclusive table lock for a long time, preventing the poller writes.
The other "work-around" idea would be to increase the time before the hot-standby activated. Right click the Hot-Standby server in Monitor Polling Engines and use Configure Hot-Standby. Default value is two minutes.
Also, check that the time is synced off the same source for each server.
See if you are losing data during that time.
We have to run our own Rebuild Index jobs to keep our databse clean. Everytime we run them we get the failover message. We don't get failover for the Solarwinds nightly maintenance. Since we don't lose any data so we just ignore it.
I wish Solarwinds had chosen a different method for the HSB to check the status of the other pollers, as this way seems a bit flaky.