How to prevent performance degradation while recovery

(Jae) #1

ElasticSearch is very awesome, but I really really want ElasticSearch to be
more robust with recovery situation.

Today, 2 elasticsearch instances were down. (don't worry, servers were
rebooted by AWS) and cluster status was gone to yellow.

The problem was, a lot of bulk index requests were dropped due to timeout
with 10s.

Is there anybody who knows how to mitigate this problem such as parameter

Would it be better if I change c.r.a parameters? If so, what would be
tuning values?

Thank you
Best, Jae


(system) #2