Liam Sutton
Liam Sutton

Reputation: 51

AWS Elasticsearch master instance connection status red for no reason

I am using a t2.micro es instance on AWS and the "master instance connection status" keeps going into red and making all calls to the instance return a 503 error. All my calls to the instance are done through AWS Lambda.

I can't find any reason for this, it's not happening at the peak times (I can see other times when there is a higher volume of requests and it isn't stopping the instance) but it's stopping all calls and breaking the site it's connected to. I also can't find any documentation on fixing this on the AWS site.

This is what I see on the status graph: A Master Instance connection graph with all green except for a large red spike in the middle

Upvotes: 2

Views: 982

Answers (1)

k00b
k00b

Reputation: 48

Given your small instance size the issue is likely low memory. When I see these errors, they coincide with GC warnings in the application logs.

[2020-11-17T11:07:42,873][WARN ][o.e.m.j.JvmGcMonitorService] [eb7abe6ce370d12f58e9dc712fe6c31a] [gc][148018] overhead, spent [1.8m] collecting in the last [1.8m]

Upvotes: 2

Related Questions