Right, we moved one of our main sites onto a new AWS box with plenty of grunt as it would allow us more control that we had before and future proof ourselves.
About a month ago we started running into issues with high memory usage at the same time every day. In the morning an export is run to export data to a file which is the FTPed to a local machine for processing. The issues were co-inciding with the rough time of the export but when we didn't run the export one day, the server still ran into the same issues. The export has been run at other times in the day since to monitor memory usage to see if it spikes. The conclusion is that the export is fine and barely touches the sides memory wise. No noticeable change in memory usage.
When the issue happens, its effect is to kill mysql and require us to restart the process. We think it might be a mysql memory issue, but might just be that mysql is just the first to feel it. Looking at the logs there is no particular query run before the memory usage hits 90%.
When it strikes at about 9:20am, the memory usage spikes from a near constant 25% to 98% and very quickly kills mysql to save itself. It usually takes about 3-4 minutes to die.
There are no cron jobs running at that time of the day and we haven't noticed a spike in traffic over the period of the issues. Any help would be massively appreciated! thanks.