SocketTimeoutException: null

[Christopher_Glantsch (Christopher Glantschnig)]

I am having a cluster on elastic.co and I have there basically 5 indexes. 1 of them has many write operations ( like 100 a minute ), where some of them are also bulk inserts.
the other indexes are mainly there for reading and there are like 50 read operations a minute, but it is just written like a few times a day.

So currently I get about 300 times a day a SocketTimeoutException: null, host not reachable Exception on my Java Client.

Any ideas what I could change in the cluster to fix it or improve it?

[Christian_Dahlqvist]

Do you have monitoring enabled? If so, what does your heap usage look like?

[Christopher_Glantsch (Christopher Glantschnig)]

that are the values of the 2 nodes we have

[Christian_Dahlqvist]

It seems heap is at 75% which is when GC kicks in. If it stays at this level you will need to either reduce heap usage or increase the size of your cluster.