In a previous blog post we had walked through a number of ways to handle multi-tenancy in Cumul.io. Now in this tutorial we will walk through how to configure your Cumul.io dashboard with multi-tenancy based on user information on login, which we manage with Auth0. Before we dive into the specifics, let me give you some context on what these two platforms are used for;
Cumul.io: SaaS that allows you to create your own data insight and analytics dashboards that you can easily integrate and embed within your own platform.
Auth0: A flexible solution to add authentication and authorisation services to your application. You can connect any application to Auth0 and define how you want your users to log in.
You can either clone this repo and follow the steps exactly, or you could take this as a guide for your own setup. The repository already implements a simple single page web app which has an integrated dashboard that you can run locally. To be able to follow these steps you will need a Cumul.io account (Free trials are available here) and an Auth0 account.
Why set up multi-tenancy on your dashboard?
Multi-tenancy on your dashboard gives you the freedom of deciding who sees what when they visit it. It lets you decide for example what an internal user may see (such as an employee) or what an end user may see.
Assume you have a company which owns various products and you want to see how sales are going for each product. You want to have a dashboard that displays this but you don’t want all employees to see all products, rather the one related to their department. This is only one simple example of why you would want user-based filtering. It is so that you can make sure an end user only gets to see what’s intended for them.
Our aim in this tutorial will be to get our imaginary employees Angelina and Brad to go from seeing this:
First step, log into your Cumul.io account and create a new dashboard (we’ve named ours “Multi-Tenancy Demo”). In this example we use the “United Widgets Sales” dataset, which you can find at: DATA → (+) → Demo Data → United Widgets Sales →Import:
We have selected a grouped line chart and have used the Month in Datetime for the X axis, the Amount for the Measure and grouped by Product name to obtain the following:
To place the chart and attach data to it, it is a simple drag and drop. You can find the chart and other options on the left ‘Add items’ tab. You can also drop data into the relevant axes from the Data tab.
Step 2 – Create an Application and Users with Auth0
Let’s assume our company has 2 employees. Angelina Julie and Brad Pots. Now we will add them as users to our app on Auth0.
In the Applications menu, create a new Application and select Single Page Web Applications:
In the user_metadata of these users, add their firstName and language. In app_metadata add their department. user_metadata is meant for user preferences that they could easily change, whereas app_metadata is for user information that an admin would control:
If you are following this tutorial with the example repo (cumulio-auth0):
Create a file called ‘.env’ in the root directory. Here, create the following 2 keys and add the KEY and TOKEN from your Cumul.io account (You can create one in your Profile settings under API Tokens): CUMULIO_API_KEY=XXX CUMULIO_API_TOKEN=XXX
In ‘server.js’ and ‘public/js/app.js’ replace ‘dashboardId’ with your own dashboard’s ID which you can find at the bottom of the Settings tab on your dashboard editor.
This step is where we allow our app to receive metadata for users once they log in. Here, we will use user_metadata for user preferences such as language selection, and we will use fields in app_metadata as parameters for the dashboard to filter on. This will allow users to only see the Sales for only their department.
In order to achieve this, we will add relevant metadata to our jwt tokens. For this, we can add a rule in Auth0:
Go to Auth Pipeline → Rules and create an empty rule with name ‘Add metadata to token’. Here you can use the following code or modify it to your needs. This rule adds the user_metadata and app_metadata to your jwt token.
⚠️ This makes sense for this use case as the user_metadata and app_metadata are both very small here. However if you have large metadata content you may want to consider making a separate call to Auth0
In the following section, we will add the parameter to the dashboard and filter the line chart on this parameter.
Step 5 – Create Parameters on the Dashboard
Now the application is receiving department as a parameter, but the dashboard is not doing anything with it. If you run (npm run start), go to localhost:3000 on a browser and log in as Brad, you will still see a chart containing sales data for all departments:
Now it’s time to create a department parameter on the dashboard and add it to a filter:
On the FILTERS tab on your dashboard, create a parameter called ‘department’ of type Hierarchy and add as many department names (from Product Names in the dataset) to this parameter as you want, or even leave it empty (in this case, you will see ‘No Data’ on the Cumul.io admin dashboard once you add the parameter to a filter in the next step). For more information about how this works and how to set it up, have a look at this article about Parameterizable filters.
Create a filter by going to the FILTERS tab and adding new filter to DASHBOARD FILTERS. You can define your filter as follows:
Once you apply this new filter, the chart you see on your Cumul.io dashboard will display data only for product names that are included in the department parameter, and nothing if you haven’t added any values to the parameter.
Now you’re all set! You can run npm run start and go to localhost:3000 on a browser. Once you login as one of your users you will see that they can only see data related to their department.
Step 6 – Add Extra Metadata (Optional)
Now that you’ve filtered your dashboard depending on what department your employee works on, let’s walk through what you can do to add something else if you want to. This section is a summary of steps you will need to take, and can be adjusted as you wish.
Assume you don’t only want to filter based on department, but you also want to filter a dashboard based on an employee’s date of employment. For some reason, let’s say you don’t want an employee to see sales data from before they joined. Here’s what you can do:
Let’s add a ‘join_date’ field to our users’ app_metadata on Auth0. For example: