Next to triggering an alarm when storage utilization reaches a certain threshold, I am looking for a solution to also trigger an alarm when the amount of data written in a certain time frame is more than X GB, for example more than 100 GB in the last two hours (sliding window).
Another idea is to also trigger an alarm when no data was written in a certain time frame.
Is there a simple way to trigger those alarms using standard DataMiner alarm template functionality?
I believe that this can be accomplished by monitoring the FREE DISK SPACE metric (for a Microsoft OS server – there might be a different metric to indicate the disk space usage for other products such as a LINUX OS, or a a specific storage product) and configuring an ABSOLUTE alarm threshold (in the alarm threshold there is a drop down where you can choose between NORMAL/RELATIVE/ABSOLUTE/RATE. If you do that, your normal value will become [BASELINE] and you can click on that to define it. In your use case you would be using the option “Automatically update the baseline” and set it to 1 day. This means that the normal level is redefined each day by DataMiner, and it will trigger an alarm if the actual value deviates from that normal value more than what you indicated in the alarm threshold (e.g. if you put 1000 MB as your threshold, if then the disk space changes for more than 1000 MB in 24 hours, the alarm will trigger). The RELATIVE threshold allows you to do the same, but expresses the deviation as a % of the normal value instead of an absolute deviation. Note: trending has to be enabled on that metric, so that DataMiner can look up the historical value when it updates the baseline. If you want to trigger an alarm on both a fixed threshold AND on a variation, simply hover over the metric in the Alarm Template Editor, and click on plus to duplicate the metric. And then you can set both the fixed threshold as well as your absolute variation. More information is also available in the DataMiner Help if you are interested in more details:
DataMiner Help
Another option, that can also be used for storage clusters is to create an aggregation rule on a view containing the Microsoft or Linux elements. For Microsoft platform elements, you can aggregate the free space of either a specific disk or all disks. By running the aggregation every 2 hours and adding a rate alarm for a difference of 100GB, you can get an alarm if over 100GB was written. Similarly, you can set up an alarm for no data or, as there is likely to be some disk activity e.g. less then 10MB written to disk.