Problem Definition

Every server maintenance with a period above approximately 10min will cause all the agents to be disconnected. This is a big pain when we have multiple agents since we have to restart them one by one back

Suggested Solution

When an agent is installed, the defaults propperties affecting the reconnection are:

wrapper.successful_invocation_time=3
wrapper.max_failed_invocations=5 

^{Reference: Remote agent does not restart after server outage - Atlassian Documentation}

Please set the wrapper.max_failed_invocations parameter by default with a bigger value like 1440 (for 2h server downtime or even a bigger value):

wrapper.max_failed_invocations=1440

This will save the time of Bamboo admins with higher agents tiers, avoiding the need to restart the agents after server maintenance bigger than 10min.

Workaround

Manually change the setting after installing the agent as described in Remote agent does not restart after server outage - Atlassian Documentation