Elasticsearch Indexing Issues

We have a 5-node cluster, currently running ES 5.6.11. When there's an increased load in indexing, heap usage reaches 90% in one of the nodes, while some have only 40-50% heap usage. Because of this, elasticsearch takes 70-80 seconds sometimes to index even 5-10 documents. We don't make the thread sleep after every bulk indexing. I am attaching a screenshot of heap, CPU & disk usage. Is there any way we can mitigate this issue? We have CMS GC as that is the only one supported in ES 5.6.11.

Is there any way to make the indexing faster? Our search is working just fine, as due to the index heavy load, we have increased refresh_rate to 2s. Only indexing is causing issues now.

PS: 2nd Row is the current Master Node, with 75% heap usage

Please note that version is very much EOL and no longer supported, you should be looking to upgrade as a matter of urgency. As such you are unlikely to get a resolution here other than to upgrade.