[Feature request] Add possibility to load further aggregator configurations after node was started

[BEG-Katzenberger]

Currently the aggregator node only reads it's configuration once right after start and it is not possible to update this config later.

Our use-case: we have two independent systems with separated aggregator configs and the aggregator node is started by the first running system. If the second system is started the aggregator node on the first system should now take over the aggregator configuration from the second system.

Unfortunately running an other aggregator node on the second system leads to a toggling view in the rqt_robot_monitor GUI.

[ct2034]

Thanks @BEG-Katzenberger for the feature request. PRs are welcome. ;-)

Do you think this could also be solved by the node lifecycle?
I am currently trying to think of different ways to make your use case possible, but I guess there are multiple options:

• Kill original node and make new one with new config
• Make config expandable at runtime
• Decouple separate aggregators completely such that they are separate processes, anyway.

What do you think?