0

I have a old java application running on an AWS EC2 but a couple of times per week the application is very slow or down and I have to restart the system which is very bad for our clients. It's an old application and it's using a lot of the local file system and also has some cron jobs what Is a big constrain for adding a load balancer because we can't have multiple instances running behind a load balancer. The instance is not immutable. I know the solution is to remove these this local filesystem and maybe move the cron jobs to lambdas so that I can have multiple instance running the same application, but that will take time to achieve it.

Does anyone recommend a quick solution (even if temporary) that I could apply now to have a more stable or available system, so that I gain time to do the right changes? I'm already using a m1.xlarge instance.

I'm also already using new relic to monitor the system but it seems the slow transactions and errors that happen when system is very slow are more a consequence than a cause.

Thank you! Any help would be appreciated!

dbeja
  • 101
  • M1 is a very old instance type, M4 is faster for the same money. You need to work out what the problem is before we can really help, it could be a memory leak or similar. Since the OS is probably fine this is down to your application. I think you need to fix the application and design it for the cloud. – Tim Mar 17 '17 at 21:10
  • Thanks. I would like to upgrade to m4 but the region where I'm running the app doesn't have m4. But maybe it would be a good start to migrate ec2 and ebs to a region that has m4. – dbeja Mar 17 '17 at 21:55
  • m3 is still a good step up. Regions choice is usually about latency and compliance, but technically moving is easy. – Tim Mar 17 '17 at 23:00

0 Answers0