Create an alert based on the disk consumption in the last 2 hours.

[andresjavier]

Hi Community!. 

I've a case where the disk D:\ of a bunch of Citrix Servers are being filled up quickly and sometimes when we're received the alert of disk usage it's too late and the services stop responding. 

Then, we would like to setup an alert based on the % of disk usage in the last 2 hours. For example, if the disk usage increase in 40% in the last 2 hours trigger an alert. 

They already have a similar alert configured in splunk, but I would like to do the same in Dynatrace. 

See below the splunk alert currently in place:

index=perfmon host=jwppxdsh* sourcetype="Perfmon:FreeDiskSpace" instance="C:" OR instance="D:" counter="% Free Space" instance!=_* instance!=Hard* earliest=-2h

| stats min(Value) AS min,first(Value) AS first by host,instance

| eval Percent-Size-Change=first-min

| search Percent-Size-Change>=40

[AntonPineiro · Accepted solution]

Hi,

You can use new Disk Edge alerting.

Instead 2 hours, set how many 10 seconds window should be under your threeshold.

Best regards

[andresjavier]

Thanks @AntonPineiro ... I'll try also with this approach.!

[michal_lewi · Accepted solution]

Hello @andresjavier,

you can achieve this result by creating custom metric selector:

builtin:host.disk.free:splitBy():sort(value(auto,descending)):timeshift(-2h)-builtin:host.disk.free:splitBy():sort(value(auto,descending))
and create custom alarm based on it:

you can also check new feature called “Disk Edge” which allows you to better evaluate disk performance.
Settings API - Anomaly detection for infrastructure- Disk Edge schema table - Dynatrace Docs
Best Regards
Michał

[andresjavier]

Thanks @michal_lewi .... I've modified you query a little bit to include just the 😧 Drive and it seems to be working... I'll configure the alerts and see how it goes!...