Hi!

This PR fixes two issues:

1. we noticed that sometimes we could see some log records repeated, like so:

   Processed: 500/407210 (took 0.84s)
   Processed: 26000/407210 (took 0.56s)
   Processed: 53500/407210 (took 0.51s)
   Processed: 53500/407210 (took 0.51s) <- double
   Processed: 98500/407210 (took 0.67s)
   Processed: 123500/407210 (took 0.53s)
   Processed: 123500/407210 (took 0.53s)
   Processed: 123500/407210 (took 0.53s) <- triple
   Processed: 152000/407210 (took 0.47s)
   

   it was caused by the pending key not being set correctly, so there was some sort of race condition;
2. because of the race condition we could get ThrottlingException, get logs calls have a hard limit, and they don't qualify for a limit increase in AWS, so having two, three... tasks running simultaneously could cause the error quickly, the easiest way just to ignore it as it do not affect anything.

I've also added configuration for durationBetweenPolls and timeoutBeforeFirstLogs