Reliable and robust networking

[orduno]

Feature request

Feature description

Expanding what we have in #592 on the topic of networking reliability, let's discuss what we can do to improve in this area. Here's some to start with:

• Check for timeouts at every service/action call and fail gracefully
• Verify we are receiving messages at a reasonable frequency

Please feel free to edit the ticket and add your ideas.

[mhpanah]

@orduno This is related to the first two items of #953.

[SteveMacenski]

This now covers doing the testing and service/action timeout - TF has been taken care of via #978

[naiveHobo]

Service timeout is already implemented in the check_future method which is called in tick():
https://github.com/ros-planning/navigation2/blob/fc762f659d98ce4ca2ebc5d87e6fc556742ffc9a/nav2_behavior_tree/include/nav2_behavior_tree/bt_service_node.hpp#L98-L114

[SteveMacenski]

OK, I know a few are going to use this, but we need a survey of all of them to make sure that all of them do / will have a timeout on the spin_until_future_complete()s -- or are using the asynchronous callback methods which I think I only did in the BT action node. You'll find service callers in waypoint navigator and recovery server too.

[SteveMacenski]

@naiveHobo you get a chance to finish this and there was no action?

[SteveMacenski]

Pinging again @naiveHobo

[SteveMacenski]

Things requiring timeouts (in spin until future complete)

• nav2_behavior_tree/include/nav2_behavior_tree/bt_action_node.hpp
• nav2_lifecycle_manager/src/lifecycle_manager_client.cpp
• nav2_bringup/nav2_gazebo_spawner/nav2_gazebo_spawner/nav2_gazebo_spawner.py
• nav2_rviz_plugins/src/nav2_panel.cpp

After which this ticket is complete - pretty small project actually :-)