What to use as Data Source credentials?

[ovonel]

Hi,

What is the standard for Data Source credentials?

Should I use my own credentials? what if I leave the organization??

Furthermore, I do not understand why it says Off in Scheduled refresh:

As it can be seen in the 'Refresh history' there is clearly a schedule...

[edhans]

You can use your own credentials, or work with the IT department to create a service account that can be used.

Your refresh failed 4 times. When it fails the fourth time, the scheduled refresh is turned off to reduce load on the service. You need to fix the issue then turn it back on. If you turn on Scheduled Refresh without fixing it, when it fails again the 4th time, MS will turn scheduled refresh off again.

[arock-well]

@edhans What about ovonel's other question:

Should I use my own credentials? what if I leave the organization?

[edhans]

When you set up the alerts for failures, you should always add an admin or someone on general that will know this.

Using your own credentials is fine, and when you leave, it will fail when your account is locked down, but the alert will go out and can be fixed.

The problem with a service account as I mentioned is when you do leave, that password must be changed too, and now that will impact MANY more refreshes than just yours.

In a true enterprise environment, the Power BI admin team should manage refreshes in the production environment of a deployment pipeline to mitigate this. In a smaller org, less feasible.

[arock-well]

Is there a way to easily see what accounts are being used to refresh datasets, dataflows, etc.?

[edhans · Accepted solution]

You would need to use the Power BI Rest API to get this info. Admin - Get Refreshables - REST API (Power BI Power BI REST APIs) | Microsoft Learn for example will give you info on who has configured a dataset to refresh.

Dataflows - Get Dataflow Data Sources - REST API (Power BI Power BI REST APIs) | Microsoft Learn will show you what datasources are used in a dataflow, and there is a sister command for datasets.

There isn't one call that gets all datasets/dataflows and returns who has access to those data sources, but it isn't quite that simple as any data source that is through a gateway, the end user configuring the refresh only needs access to the gateway and doesn't configure each data source, so even for data sources that don't require a gateway, by putting them in the gateway, it is much easier to manage. The permissions are granted by the gateway, not the user configuring the refresh. They just need permissions as a gateway user.

But through the few Rest API's I've shown and the dozens and dozens of others, you could put together the info you need.