DataMiner reports on the active users and also provides some logging for user activity. However, it is not clear if there is a way to report on historical user activity.
Use Case
- Ability to track the amount of active users at a specific point in time
- Since the amount of concurrent users accessing the system may impact the performance of DataMiner, it would be great to have some data on historical user activity that could be correlated with periods of degraded performance
From the above, it seems possible that we can create a script or connector that gathers user activity. However, the use case is more towards something generic.
Hey Rene,
sounds like something we would want to cover with the upcoming GQI. Not yet on the current version, but we are including queries on both alarm events and information events (and user activity is exposed via information events). And I would have to look at some of the details here, because we have an event when a user logs on and off, but your use case is about knowing how many are online at a given point in time and that would require some data manipulations (and the GQI allows you to do many manipulations already - just not sure if we would have everything to cover this one specifically). Of course you can also keep it a little more simple and just look at the nbr of connect and disconnection events, that gives you also a sense of user activity.
In any case, the GQI really promises to be a real game changer for any sort of data query. I already had a lot of fun just playing around with it.
For now, for your use case, I would just load the history of the information events related to user activity in the alarm console for the period of interest, and copy that into excel and see if you can do a quick pivot table to whip up the graph that you want that shows the nbr of online users over time. I guess that can work for now, because you would not do that every day probably. Moreover, once you have the Excel, you can easily use it as a template to copy new data into it from another timespan later.
PS: nbr of online users is not really going to give you a sense of the load, because DataMiner can deal with a very high nbr of users if they are regular operators (i.e. it would only be relevant if those users are doing a lot of heavy queries).
Hi,
An addendum to the above.
Currently we’re working on a new feature in the core of DataMiner called the SPI Framework.
The goal of said feature is to easily expose System Performance Indicators. Allowing for in-depth monitoring and analysis of the DataMiner.
As part of the SPIs that we’ve planned, we would keep track of the number of active connections and the rate of traffic over each connection.
Once this SPI is available we can hook this up to the GQI with little effort to allow for the use case of the initial question.
There are 2 important notes regarding the SPI Framework to keep in mind.
– Not all SPIs will be available with the launch of the feature, this is something we will expand on over time. Starting off with the SPIs that we have seen the highest need for at first.
– Not all SPIs will be enabled by default. As always in software development, everything has a trade off. Thus many of the SPIs will need to be explicitly enabled by the Administrator(s) of the system before they’re available in GQI. We’ve gone through great efforts to ensure that each SPI update introduces as little overhead and resource usage as possible. Similar to how we approach trending, we wish to express consideration in the SPIs that people opt for. As enabling all of them could quickly lead to an information flood that one can’t keep up with.