This allows users to monitor Rafka for errors that are not propagated to
the clients. Until now such errors were just being logged. It allows for
simple alerting rules (eg. alert when producer errors are more than 10).
We used the HGETALL Redis command, since the stats structure can be
considered a hash with stat names as keys and counters as values.
This allows users to monitor Rafka for errors that are not propagated to the clients. Until now such errors were just being logged. It allows for simple alerting rules (eg. alert when producer errors are more than 10).
We used the HGETALL Redis command, since the stats structure can be considered a hash with stat names as keys and counters as values.
Part of #28