Maintain dynamic out-of-service BMCs list and exclude them from subscription

[ajmerachandrika]

In the current design, all the BMC IPs that the redfish listener subscribes to are consumed from a file at initialization, and the application connects to all the BMCs and registers for the RAS events as a listener. If any connection requests fail, the listener keeps periodically trying to reconnect and this can lead to logs overflowing with failure notifications, in case the node has been taken out for service.

This PR provides a way to remove these nodes dynamically from the list of IPs that the listener should try connecting to.

1. Users can enable/disable monitoring on a node by sending request on REST Endpoint /api/nodes/{node-ip}. This information will be persisted to handle the case of application restart. It will be stored as a file to the local file system.

EXAMPLE:

CURL command to disable monitoring on a node:

curl -X PATCH \ 
  http://localhost:9000/api/nodes/10.1.1.1 \ 
  -H 'Content-Type: application/json' \ 
  -d '{"monitoring-enabled": false}'  

CURL command to enable monitoring on a node:

curl -X PATCH \ 
  http://localhost:9000/api/nodes/10.1.1.1 \ 
  -H 'Content-Type: application/json' \ 
  -d '{"monitoring-enabled": true}' 

In case there are two instances of the Redfish listener running for redundancy, the CURL commands will have to be run against both the instances as the file will not be synced across the instances by the application.

2. Users can also fetch the nodes being monitored lists using REST Endpoint /api/nodes.

CURL command to display the nodes being monitored

curl -X GET \ 
  http://localhost:9000/api/nodes