Hi,

 

Is there an attachment point for alerts to be triggered from the metric of I/O Latency?

 

Reason for asking is that I could do with an alert checker which would have spotted this issue (the scale is in seconds!):

 

 

It went unnoticed overnight until it crippled a virtual host when users started accessing VMs on it this morning. The underlying fault was a RAID card issue, but nothing else detected it and I’ve just found that Observium ‘knew’ about it, but we have not alerts which would have caught this.

 

Any ideas?

 

Cheers!