Identity

At first glance, you might wonder why identity belongs in the SaaS story. It’s true that there are any number of different identity solutions that you can use to construct your SaaS solution. You could even suggest that your identity provider belongs somehow outside the scope of our control plane discussion. However, it turns out that multi-tenancy and the control plane often have a pretty tight binding to your SaaS architecture. The diagram in Figure 2-5 provides a simplified view of how identity is applied in multi-tenant environments.

Figure 2-5. Binding users to a tenant identity

On the left you’ll see the classic notion of user identity that is typically associated with authentication and authorization. It’s true that our SaaS user will authenticate against our SaaS system. However, in a multi-tenant environment, being able to authenticate a user is not enough. A SaaS system must know who you are as a user and it must also be able to bind that user to a tenant. In fact, every user that is logged into our system must be attached in some way to a tenant.

This user/tenant binding ends up adding a wrinkle to our system’s overall identity experience, requiring architects and builders to develop strategies for binding these two concepts in a way that still conforms to the requirements of your overall authentication model. This gets even more complicated when we start thinking about how we might support federated identity models in multi-tenant environments. We’ll see that, the more the identity experience moves outside of our control, the more complex and challenging it becomes to support this binding between users and tenants. In some cases, you may find yourself introducing constructs to stitch these two concepts together.

When we dig into onboarding and identity in Chapter 4, you’ll get a better sense of the key role identity plays in the broader multi-tenant story. Getting identity right is essential to building out a crisp and efficient strategy for introducing tenants into your SaaS architecture. The policies and patterns you apply here will have a cascading impact across many of the moving parts of your design and implementation.

Metrics

When your application is running in a multi-tenant model, it becomes more difficult to create a clear picture of how your tenants are using your system. If you’re sharing infrastructure, for example, it’s very hard to know which tenants are currently consuming that infrastructure and how the activity of individual tenants might be impacting the scale, performance, and availability of your solution. The population of tenants that are using your system may also be constantly changing. New tenants may be added. Existing tenants might be leaving. This can make operating and supporting multi-tenant environments particularly challenging.

These factors make it especially important for SaaS companies to invest in building out a rich metrics and analytics experience as part of their control plane. The goal here is to create a centralized hub for capturing and aggregating tenant activity that allows teams to monitor and analyze the usage and consumption profile of individual tenants.

The role of metrics here is very wide. The data collected will be used in an operational context, allowing teams to measure and troubleshoot the health of the system. Product owners might use this data to assess the consumption of specific features. Customer success teams might use this data to measure a new customer’s time to value. The idea here is that successful SaaS teams will use this data to drive the business, operational, and technology success of their SaaS offering.

You can imagine how metrics will impact the architecture and implementation of many of the moving parts of your multi-tenant system. Microservice developers will need to think about how and where they’ll add metrics instrumentation. Infrastructure teams will need to decide how and where they’ll surface infrastructure activity. The business will need to weigh in and help capture the metrics that can measure the customer experience. These are just a few examples from a long list of areas where metrics might influence your implementation.

The tenant must be at the center of this metrics strategy. Having data on consumption and activity has significantly less value if it cannot be filtered, analyzed, and viewed through the lens of individual tenants.