I made an app for video and audio transcoding using node.js, PostgreSQL, rabbitmq, EC2 with EBS, and Kubernetes.
I am a little bit worried about storage size and computing power. Computing power is easy to solve but storage is more problematic. The best will be a solution that automatically scales storage size but I am not sure if Kubernetes supports that.

Right now I am thinking of swapping EC2 to bare metal with Hadoop hdfs but I feel it is overkill.

The second idea is to buy a few instances of ec2 and for every instance add own EBS, but again I am not sure if Kubernetes supports such think, at the end every EC2 instance need access to another EC2 EBS storage.

What do you think? Maybe there is an easier and cheaper way.

CodePudding user response：

So one thing you can consider since you are on AWS is to use EFS (elastic file system) as a volume type. By adding the EFS provisioner and EFS storage class to your cluster, you can make use the dynamic sizing that EFS gives you. you wont have to worry about running out of space.

You can check out the efs provsioner here!!

I have been using this storage class for other applications that are not too intense on read/writes and have had a lot of success with it! You also can mount the volume across multiple AZs so you get HA (high availability) as a plus!