Revision history [back]

For something like this a LifeCycle node would fit quite perfect. The node is written in Cpp but only uses a standard rclcpp::Node as base interface.

With a LifeCycle node you could easily restart your node from outside and maybe fiddle with the parameters in between startup tests.. (or try to reset your external devices etc. pp.)

But this needs to be implemented directly into the node.

How you could subscribe to the logging level is also open. I beleive that you could hook into the rclcpp logging system and listen to the output from your given node. If you then register error output you could toggle a LifeCycle "restart" (go from active to unconfigured and back)

But again, this is not something you just enable. It needs to be done in code. But I'm sure that the maintainer would like to welcome this addition.

Also be aware that you are working with a foxy node and tagged galactic. This could easily work but also come up with some nasty runtime errors that are not already known/fixed by the maintainer. Maybe considering using foxy to be safe or help identify upgradeable points for the node to be 100% galactic ready. There are exact changes described for nodes upgrading from foxy to galactic. Maybe start searching there if you want to give something back.. :)