cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

DMI Report: How can I report on the volume of users over the last 12 months?

mick_huynh
Inactive
 
7 REPLIES 7

cosmin_gherghel
Dynatrace Pro
Dynatrace Pro

Hi Mick,

Can you provide a bit more information on what you want to report on? Would you like to know number of unique users? Bytes transfered from all users?

mick_huynh
Inactive

Yes. The number of unique users.

Similar to the this chart but without the affected users line. Viewed on a weekly/monthly period.

cosmin_gherghel
Dynatrace Pro
Dynatrace Pro

Mick,

To create a similar chart, create a new report with a chart section including the time dimension and any other dimension you want to report on (set these to filter or column) and then add the unique users metric. You can use the chart settings tab to change the way the chart looks.

Take a look at the DMI documentation which contains a lot more information around creating such reports. https://community.dynatrace.com/community/display/...

mick_huynh
Inactive

Hi,

I have attempted to create a DMI as you have suggested but when I select time period beyond the last 30days, I no longer have the Unique users metric.

The reason why unique users metric is not available for historical time ranges is that it would require DCRUM to keep a data for all those users for a very long time. The database size and the time to render such a report would grow significantly.

It could be an idea for extending product capabilities, but even to keep data with just one day resolution for historical time ranges we would first require to identify what kind of aggregates are required and keep information for those aggregates only.

For other resolutions it would become virtually impossible to properly compute unique users. Example: unique users values during the week were [Mon: 10, Tue: 12, Wed: 20: Thr: 14, Fri: 23, Sat: 10, Sun: 2]. What would be the unique users count for the whole week? We can only tell it would be something between 23 (everyone who visited the site this week also visited the site on Friday) and 91 (each user never returned to the site during this week).

The solution could be to use Splunk or Elastic integration to export data. It should work for one day resolution. For other resolutions to work you would need to export raw data for aggregates that you are interested in.

mick_huynh
Inactive

Sebastian,

Unique users may be the wrong metric for what we are trying to do. We just need the volume of users.

If a users comes in multiple times a day, week, or month, we need to count him/her to the total anyways.

I was just wondering if "User sessions" metric wouldn't do the trick?

User sessions = "The number of user HTTP sessions. The count can be identified by information contained in intercepted HTTP cookies or by HTTP authorization."