What could be a reason for the cluster taking so long before an old captain is downgraded?

[FreekMulders]

In my cluster with 3 peers I brought down an availability zone at 10:05 containing the shc_captain.
The other two members were working fine.
At around 10:30 a new captain was elected and my old captain was downgraded.

If I bring down the single captain, a new captain is elected within 2 minutes.

What could be a reason for my cluster taking so long before an old captain is downgraded?

[codebuilder]

How long was the captain down? If you were running on two clustered search heads only for any period of time it is possible that the cluster encountered a split-brain situation. With only two nodes, a quorum can't be reached so neither is able to (quickly) decide/vote on who should be captain.