About the use of this diagnostic node package about diagnostics HOT 6 CLOSED

Yes, the get_node_names() (https://docs.ros2.org/galactic/api/rclpy/api/node.html#rclpy.node.Node.get_node_names) method should return all running nodes. I am not sure, though, how quickly a node is removed after it dies. Maybe you have to test this.

from diagnostics.

You can publish the results similar to how it is done here https://github.com/ros/diagnostics/blob/ros2/diagnostic_updater/diagnostic_updater/example.py

from diagnostics.

Hi @TZECHIN6. I agree that this may be a common use case. Currently, there is no pre-made node for that. Please feel free to add one to: https://github.com/ros/diagnostics/tree/ros2/diagnostic_common_diagnostics :)

from diagnostics.

To verify a node is dead or not, what is the best way to do so, a very simple method is to publish a heartbeat message say every single second, but this come to creating a whole bunch of messages if there are many nodes.

is there any build in function in ros2 can directly access its state?

get_node_names() Function in rclpy will list out all node within ROS2 environment, I am not sure does a node can be seen mean that node is alive and functioning…

from diagnostics.

I can come up with a very quick way to test it using python list element check, but this come down to hardcoding the node name needed to be checked... which some time the launch file can rename the node. So I am not so sure is there any dynamic way to solve this if using get_node_names().

I can share it to the common diagnostics, do you expect the output of the diagnostics result in a form of this /diagnostic topic? Acutally I havent use this before, so have to understand the structure of it..

from diagnostics.

I just spotted that there is a class (here) actually doing something to what i want, check a node is alive or not. but I am not so sure how to use it. I will rise another issue about that. Thanks

from diagnostics.