On AWS, I run Elasticsearch 7.3.0 cluster composed of 2 t3.small nodes and 1 t3.small node. Each node has 16GB EBS volume for root volume. Nodes are managed by ECK.
Is master node in abnormal state for some reasom, or just under heavy load constantly, so I'll need to give more capacity to root volume? I don't know how to debug it.
I'm talking about burst credits of root EBS volume, not CPU credits. I have separate EBS volume on each node, and its status looks okay even after accident.
After the burst balance goes to zero, ES pods on master node keeps restarting because of slowness caused by it.