Elasticsearch gc overhead

[Akaren]

I am using ES 7.1.1. I have 4 big indices (around 20-40GB, 2 shards) and 6 small indices (500MB).
Disk memory is 32GB, heap size is set (min and max) to 13GB. And only one node.

Searching is so slow and I can see warnings:

   [WARN ][o.e.m.j.JvmGcMonitorService] [] [gc][494435] overhead, spent [13.7s] collecting in the 
last [14.6s] [2020-02-22T12:08:58,291][WARN ][o.e.m.j.JvmGcMonitorService] [] [gc][young] 
[494438][79303] duration [5.3s], collections [1]/[6.1s], total [5.3s]/[1h], memory [8.2gb]-> 
[8.2gb]/[12.3gb ], all_pools {[young] [90.7mb]->[16mb]/[865.3mb]}{[survivor] [84.6mb]-> 
[107.3mb]/[108.1mb]}{[old] [8gb]->[8.1gb]/[11.4gb]} [2020-02-22T12:08:58,292][WARN ] 
[o.e.m.j.JvmGcMonitorService] [] [gc][494438] overhead, spent [5.3s] collecting in the last [6.1s] 
 [2020-02-22T12:10:00,624][WARN ][o.e.m.j.JvmGcMonitorService] [] [gc][young][494493][79304] 
duration [7.3s], collections [1]/[8.3s], total [7.3s]/[1h], memory [9gb]->[8.4gb]/[12.3gb], all_pools 
{[young] [858.2mb]->[10.3mb]/[865.3mb]}{[survivor] [107.3mb]->[108.1mb]/[108.1mb]}{[old] 
[8.1gb]->[8.3gb]/[11.4gb]}

Any idea what to do?