Several times a day we receive a Linode Disk IO Alert (which is set at an average of 1000 Ops/sec last 2 hours).

Current research:

The SWAP IO doesn't coincide with all the IO peaks:

The memory usage is around 60% while a peak is occurring

It doesn't seem to be an application issue neither, as the IO peaks aren't related to higher network traffic


I had the luck to coincide with one of these peaks, and with iotop I could realize all the activity was due to Apache read operations.


Our configuration is the following:
Ubuntu 12.04 Intel(R) Xeon(R) CPU L5520 @ 2.27GHz
1GB RAM
256MB SWAP
Main technology: LAMP, Zend, Redis and Node
Apache version: 2.4.6

Overall, there are a couple of questions:
- Why apache would read that amount of information?
- Could this situation be normal, and we can just update Linode disk IO alert threshold?

You can adjust the alert settings (in your Linode manager click on your Linode and then choose the 'Settings' tab).

There may be an issue with your site(s) if they are causing sever IO spikes. What kind of sites are you running? The more you tell us about your Linode/websites the easier it will be for us to help you.