As I mentioned, I know how to reduce heap size for Elasticsearch.
On my current system, Elasticsearch is not taking up 50% of my memory (128GB in total, so it’s consuming about 25%). On the old system, it was taking up 100% including swap space which is why performance suffered greatly.
Hence my questions, specifically what the consequences of reducing heap size to 2G or 4G will be, if any.
No other problems, just Elasticsearch taking up a lot of memory.
But since you mentioned it, can you give any specific recommendations?
- WEB_CONCURRENCY - does this need to be changed with only 5 users?
- ZAMMAD_SESSION_JOBS_CONCURRENT - as I understand this only becomes relevant if there are 40 agents or more.
- ZAMMAD_PROCESS_DELAYED_JOBS_WORKERS - same as above.
- ZAMMAD_PROCESS_SCHEDULED_JOBS_WORKERS - I only have one single scheduled job in Zammad so I assume this won’t have any effect.
- REDIS_URL - it is my understanding that this is default anyway with Zammad 6+, so nothing to do here.