We were running logstash 2.4 with 128MB and 256MB heap since ~2 years and it is running very fine. We are using input plugin as udp with port, then a grok fliter and output plugin as Elasticsearch of same version as logstash.
Now we are upgrading it to version 6.2 and seen it is failing to start with the error it require more memory.
We started with 512MB RAM as heap and then increase the same to 1GB, but in both cases logstatsh process starts but kills automatically after some time ~10 mins duration with the error of OOM.
As we are running ~60 logstash process as a cluster, we can't assign every process as 1GB. After some debugging we found there are memory queue used, which may cause this, but didn't found any particular solution to fix it.
Can some one please tell us why it is killing more frequently and is there any way to reduce the memory usage.