When running replays we often want to target specific rules. When using paced replay (using the --replay-speed parameter), this means we're redirecting the replay into a state machine with a dispatch state sending the event back on the bus. At this point we lose the previously set targeted replay and the event will run through all matching rules on the bus.
If you omit the --replay-speed this is not an issue.
This is also not a problem for evb replay-dead-letter since the target is always Lambda for those and it's straight forward to create temporary permissions fo rthat one use case.
However for non-dead-letter targets, we'd need to duplicate all targets behind the specified rules along with temporary permissions and policies for them. This is time consuming, so I'm raising this issue as a step 2 implementation
When running replays we often want to target specific rules. When using paced replay (using the
--replay-speedparameter), this means we're redirecting the replay into a state machine with a dispatch state sending the event back on the bus. At this point we lose the previously set targeted replay and the event will run through all matching rules on the bus.If you omit the --replay-speed this is not an issue.
This is also not a problem for
evb replay-dead-lettersince the target is always Lambda for those and it's straight forward to create temporary permissions fo rthat one use case.However for non-dead-letter targets, we'd need to duplicate all targets behind the specified rules along with temporary permissions and policies for them. This is time consuming, so I'm raising this issue as a step 2 implementation