I've deployed ws sockets on Amazon EC2 t3a.small instance. When I start/restarts the sockets it works perfectly fine for sometime. But after few minutes when I try to connect to the sockets it's taking more then a minute to connect and sometimes trigger timeout.
So I changed the instance to t3a.medium and now it's working pretty well. And I'm not seeing any delay in connecting/timeout to the sockets even after couple of hours.
When I check the performance of both the instances the RAM/CAPU usage is not that much. RAM usage is around ~300mb and also not reaching 100% CPU usage for both cores. So I'm not sure what could be the issue.
I don't think updating the instance type is right and scalable solution , As traffic increases I assume I would face the same issue ?
I want to know what is the bottleneck and what configuration I need to set in EC2 which is scalable ?
Update:
It seems the instance size is not an issue. Even after changing it to t3a.medium it slowed down after sometime. So figuring out what could be the issue.
Apparently there was an issue with the Apache2 setting which was allowing maximum 150 concurrent connections.
So I adjusted Apache httpd to allow for more concurrent connections.
<IfModule mpm_event_module>
StartServers 10
MinSpareThreads 25
MaxSpareThreads 750
ThreadLimit 1000
ThreadsPerChild 750
# MaxRequestWorkers aka MaxClients => ServerLimit *ThreadsPerChild
MaxRequestWorkers 15000
MaxConnectionsPerChild 0
ServerLimit 20
ThreadStackSize 524288
</IfModule>
Please note you may also need to change max open files of OS. By default it's 1024 in Ubuntu. Set it to MAX (65536)