Improvements we would like to make to the plugin and the related components.
- Extend the poller configuration -> provide a definition for our node in a new package?
- Poller should be able to add service monitors/entries to packages during a reload, and have these monitored
- on existing nodes
- Re-compile and re-deploy new code on the fly without having to * Example where it can be used: collect a new metric, and add a new graph def
- Add support for multiple controllers, or clusters of controllers
- Add support for controller clusters (round-robin rest requests?)
- Structure things in a way to allow for multiple controllers to be running different versions
- Remove assumptions from support augmentations
- We need a working deep dive tool with this environment ** Ideally using IPFIX with both ingress/egress stats
- We should be able to trigger a threshold alarm and make it appear on the link
- Should be able to do the same with Minion
- I should be able to easily route my own RPC and Sink calls
- One bundle exposes the server
- Another bundle exposes the client
- Implement using gRPC -> HTTP2 -> Java 9+
- Show how to use the situation feedback features
feature:install opennms-situation-feedback
- Use the feedback for training - see xref:training_with_ludwig.adoc[Training with Ludwig].
- Use for performance testing - create lab with 5k switches and 50k hosts
- Test with https://maxinet.github.io/
- How
- Use only async calls
- Use more specific queries
- We should be expose statistics and faults for the engine itself
- [Fault] Can't connect to ODL server
- [Perf] ODL Statistics via JMX