Update the relevant READMEs

diagnostic_aggregator/README.md

@@ -132,6 +132,33 @@ You can launch the `aggregator_node` like this (see [example.launch.py.in](examp
     ])
 ```
 
+You can add analyzers at runtime using the `add_analyzer` node like this (see [example.launch.py.in](example/example.launch.py.in)):
+```
+    add_analyzer = launch_ros.actions.Node(
+        package='diagnostic_aggregator',
+        executable='add_analyzer',
+        output='screen',
+        parameters=[add_analyzer_params_filepath])
+    return launch.LaunchDescription([
+        add_analyzer,
+    ])
+```
+This node updates the parameters of the `aggregator_node` by calling the service `/analyzers/set_parameters_atomically`.
+The `aggregator_node` will detect when a `parameter-event` has introduced new parameters to it.
+When this happens the `aggregator_node` will reload all analyzers based on its new set of parameters.
+Adding analyzers this way can be done at runtime and can be made conditional.
+
+In the example, `add_analyzer` will add an analyzer for diagnostics that are marked optional:
+``` yaml
+/**:
+  ros__parameters:
+    optional:
+      type: diagnostic_aggregator/GenericAnalyzer
+      path: Optional
+      startswith: [ '/optional' ]
+```
+This will move the `/optional/runtime/analyzer` diagnostic from the "Other" to  "Aggregation" where it will not go stale after 5 seconds and will be taken into account for the toplevel state.
+
 # Basic analyzers
 The `diagnostic_aggregator` package provides a few basic analyzers that you can use to aggregate your diagnostics.
 

diagnostic_aggregator/example/README.md

@@ -1,5 +1,7 @@
 # Aggregator Example
 
-This is a simple example to show the diagnostic_aggregator in action. It involves one python script producing dummy diagnostic data ([example_pub.py](./example_pub.py)), and one diagnostic aggregator configuration ([example.yaml](./example.yaml)) that provides analyzers aggregating it.
+This is a simple example to show the diagnostic_aggregator and add_analyzer in action. It involves one python script producing dummy diagnostic data ([example_pub.py](./example_pub.py)), one diagnostic aggregator configuration ([example_analyzers.yaml](./example_analyzers.yaml)) and one add_analyzer configuration ([example_add_analyzers.yaml](./example_add_analyzers.yaml)).
+
+The aggregator will launch and load all the analyzers listed in ([example_analyzers.yaml](./example_analyzers.yaml)). Then the aggregator will be notified that there are additional analyzers that we also want to load in ([example_add_analyzers.yaml](./example_add_analyzers.yaml)). After this reload all analyzers will be active.
 
 Run the example with `ros2 launch diagnostic_aggregator example.launch.py`