Currently, the ReconnectingCordaRPCOps does not propagate the max. number of retries specified in the corda client configuration. As a result, during a node outage or permanent network issue, the rpc client will keep retrying infinitely. This has manifested in the following 2 ways so far:
on exceptions that are not really recoverable, but haven't been excluded by accident. See: https://r3-cev.atlassian.net/browse/CORDA-3192
In the NodeExplorer: when simulating a permanent failure of the node, the explorer UI remains frozen. See: https://github.com/corda/corda/issues/5543
The ReconnectingCordaRPCOps should impose the max. number of retries provided by the configuration (via the parameter maxReconnectAttempts) and the NodeExplorer should use a reasonable number of max retries (say 3-5), so that the UI does not remain frozen for a long period.