Initialization and reconnect logic results in an endless loop

In certain situations it can happen that the initialization or reconnect logic results in an endless loop. For example, if we try to login with the sync user and this request fails, then we don't have a limit how often we should try to repeat that. And this can result in endless iterations. If there is an exception thrown, this means the logs are spammed.

All of that is using resources from Jira and also makes a bad impression on how we react to login failures. Especially because the retries are happening regularly every 30 or 60 seconds (but for each independent host + sync user).