I am trying to help a colleague with the Admin Dashboard on our instance of the Data Hub. Seems like it is not always working consistently. However, when I initially tried to run it locally with the proper application credentials I am still getting an error about global admin variables.
Is it required to be a global admin to execute the workflow? Also, I am on AP 5.3.2 and I see it says 5.3.3. would that small of a version difference cause this issue?
Thanks for any help.
Adam
Actually a more specific question now for the Monitor User Usage component. I got it working and it is pulling data but for some reason the LOGOUT data is coming back blank. These are the calls that are returning no data.
To be clear, the calls are successful (200 return) but the results are empty.
Why would the log in data be populated… but logoff not have any info? This is making it impossible for the node to have a proper data range and it can’t calculate the time that users are logged into the platform.
Any suggestions?
Hi Adam (@aehrenworth)!
Let’s have a look at both data apps separately:
The Admin Dashboard:
That was a bug on our side, which is fixed with the latest version we just uploaded. Thanks for spotting it! This runs with 5.3.2 as well, we backported it. Global Admin permissions are needed if you want to see it all and have the possibility to download a Hub snapshot at the end. Other people will “just” see what they are allowed to see, which could be sufficient for e.g. team admins.
The Monitor User Usage Data App:
You set up the keycloak service client as described here, right? I could imagine that the “Save events” part is missing and thus no data is saved. Could you double check that “Save events” is on in keycloak?
I hope that works for you! And thanks for the feedback 
Kind regards,
Lukas
I will share this my admin team to see if they can check on the keycloak set up. There is definitely actively being captured going back to February but seems like log off data for some reason specifically missing.
If it’s only the logouts that are missing, it could be that these events are not saved for some reaons, but need to be added to the “Saved types” explicitly
I am having the team check the configuration, unfortunately I don’t have access to it. If that is missing that should address these issues. In theory. thanks for the quick help.
Thanks for the help as it resolved the issues seen in my development hub. I have added the logout events and increased the expiration to 180 days in both production (previously set to 30 days)and development (previously set to 90 minutes). The results in development and production are different. For development I get the range of yesterday through today with is expanding due to the change yesterday. I can see session length and user graphs for the last two days which is an improvement. In production for the range, I get “missing”, session length is set to unknown, but the user graphs show the last 31 days (reflecting the modification made yesterday). Looking at the realm settings in both environments, they seem to be identical. What else can I check? Thanks!
Hi @mwhitma3, welcome to the community!
Great to hear that the issues got partially resolved already. Let’s see how we can debug on your production system.
Just Wednesday, we did release an improved version of the data app, with better error handling. Maybe that already helps?
If that does not help, we might look at the raw data. If you “Inspect” the job and go inside of the “User Usage Monitoring” Component, and therein to the “Get the Monitor Users’ Data” Component, the second “Joiner” should have a list of the available sessions. Does that data roughly make sense? Is there data after the “Get Logins, Refresh, Logout Table” Metanodes?
Kind regards,
Lukas
