Declutter your alarm tree

Hi there. I hope you are okay. In this post, we would like to introduce to you a new way to present alarms in your Alarm Console.

This capability will be introduced in version 10.0.12, which will be available to you early November. You will have to explicitly opt-in for this new feature to use it, and it is a system-wide setting. So, you either choose to use it across your entire DataMiner System, or you do not. We love the new capability, hence the post to draw your attention to it.

But the second reason for posting this, is because we also would love to enable this feature by default at a later version because we are convinced that it is a major improvement for everyone. 

But then again, we haven’t decided that yet, because it affects the way alarms are displayed in the Alarm Console, and that’s a very sensitive matter and we definitely do not want to cause any confusion or throw off operators unexpectedly. We will not do this without being convinced that this is the right thing to do, so you are invited to provide us with feedback in the comments below this post.

Let me explain what this is exactly about. We refer to it internally as ‘alarm squashing’. First, to be clear, it is not another way to group alarm events, for that we have our Correlation Engine and the newly released AI-powered Incident Tracking. What is it then? Alarm squashing is about optimizing how an alarm tree, which shows the entire so-called life cycle of an Alarm Event, is displayed in your Alarm Console. 

Most of you are familiar with what an alarm tree looks like, as this has not changed since day one. But if we take a step back, we might have exaggerated a bit with the level of detail we show by default in an alarm tree. Especially considering the fact that the Alarm Console is used to manage an operation in the first place, and hence needs to be focused on information relevant in that specific context.

Below a screenshot of a quite small alarm tree you could have on your DataMiner System today.

What do we see there? Firstly, this is an alarm that started as a warning, then briefly escalated to a major alarm, then dropped back to warning and eventually become normal again. This is what you see in your Alarm Console, but the alarm tree also shows a variety of other life cycle changes.

In this case, it also shows that a property on the alarm changed, it shows the element was moved to another view and it shows someone added a comment. Great information, but with alarm squashing, we aim to declutter your alarm tree and to focus what really matters in an operational context, being the severity changes. In the screenshot below, you see the same alarm tree but now with alarm squashing enabled.

Do you agree that this makes more sense for an operator? It has condensed your alarm tree to the essence, namely the severity changes. It might give a strange feeling at day one, but I’m quite convinced this improves the readability of an alarm tree massively.

Some questions that might cross your mind by now:

• Can we still see all the details that were left out in the condensed alarm tree? Yes, in the alarm card we still show the full alarm tree. 
• Are the alarm property updates or any other update then still visible in the active alarms list? Yes, we ‘inline’ replace the updated fields on the most recent visible item. So, you will for example still see if a ticket was assigned, or who the owner of the alarm is after it has been acknowledged, etc.
• What about the time field? If we update an alarm ‘inline’ (so e.g. an alarm property gets updated), the time of the alarm will not be updated. This enables you to still keep track of the time the severity changed.
• Alarm trees are limited to a maximum number of entries, which events are counted for that? DataMiner trims an alarm tree that exceeds the maximum number of entries taking into account alarm squashing. This means you will have way more meaningful information (real severity updates) presented in the tree.

For more information about alarm squashing, please refer to the dataminer docs.

That was it, I hope you are as happy as we are with this upcoming change. Feel free to activate this when DataMiner 10.0.12 is available early November. And as mentioned, any feedback is welcome!

Cheers

---