When Log Saver encounters an exception during writing log files, it stops processing log events after that.
Log Saver keeps a temporary set of log events in a local map while writing to HDFS files. After a successful write, new events from Kafka are added to this temporary map for the next batch of write. When an exception gets thrown during write, the events in the temporary map are re-tried before reading new events from Kafka.
There was a misplaced brace introduced during a refactor of code that caused the write to log file to also get skipped when the temporary map has events. This lead to Log Saver not writing any events from the temporary map, and hence the map would never get empty. Thus leading to no processing.