...
Our current design approach is as follows:
- Deprecate existing internal boolean system property: REMOVE_FROM_QUEUE_ON_EXCEPTION
- Continue to support default behavior if boolean set to false by setting # retries on receiver to -1
Create new Java API
Define callback API for senders to set callback to dispatchers
Invoke callback if batch exception occurs prior to batch removal
Implement a default callback API (see item 8 below)
Add parameters on gateway receiver factory for # retries and wait time between retries.
Modify Gfsh commands
Add option to gfsh ‘create gateway sender’ command to specify custom callback
Add options to gfsh ‘create gateway receiver’ command to set # retries and wait time between retries
Store new options in cluster config
Sender: callback implementation
Receiver: # of retries and wait time between retries
Create example implementation of Sender callback that writes event(s) and associated exceptions to a file
Security features
Define privileges needed to deploy and configure sender callback
With security, callback should only write eventId's and exceptions, i.e. no entry values should be written to disk
API Change
Risks and Unknowns
...