If you work in a large organization, you have a big decision before starting to adopt Datadog at scale: Should I use multi-org or not? The answer might be different for each company, but, taking the right one from the begging can save you a lot of work in the future. With that in mind, in today’s post, I’ll share the main factors you must take in consideration and how you can take an informed decision.

1. How is your company organized?

Which is the size of your company? Is your IT team distributed by business unit? Who is responsible for production? Who is responsible for developing the application? How is your business organized?

Those are all valid questions that you have to ask yourself for taking the decision.

2. How much correlation is relevant for your business?

Correlation is one of the strengths you gain using Datadog, the fact that you can track every step the request did along the way is a time saver and that it’s the reason you need to be very careful to not have too much orgs to the point of losing the correlation between services. As one of the advantages of multi-org is access segregation, it also means that you only can see a certain level of information of an application that is not being monitored by your org.

3. What are your company’s data boundaries?

Datadog has many mechanisms to allow you to mask sensitive data or controlling access to certain types of objects, but sometimes, you NEED to see the data.. Whether it’s for more in-depth troubleshooting, or just to understand what values you’re receiving from your integrators. 400 Bad Request, right? We’ve all been there. So, the question you need to ask yourself is: Everyone in this org can see the information it’s been stored? Is multi-org the best option to ensure full compliance with the data policy of my company?

4. Are you ready for the extra workload multi-org will create?

Some of the configurations of Datadog are not inherited from the Master account to the child’s, so you might need to replicate it for each account you have. Terraform can be your ally on this journey, but unfortunately it does not support everything (e.g. SAML), so be prepared if you decide to use multi-org.

5. Who pays the bill?

One thing that multi-org brings is a separated view of Datadog’s consumption, making it easier to identify which account is using more of your license or if you have any noise neighbor, so if you have a sophisticated show back/chargeback system in your company, having it separated can be a good way to make your FinOps team happy.

Using a single organization, you can create a custom dashboard with a view close to the multi-org one, but you must consider the extra hours for building it.

Conclusion

I know this may be disappointing if you were hoping for a simple yes-or-no answer. However, the goal of this article is to guide you in the right direction and give you the context you need to make an informed decision, and I hope it has done exactly that.

If not, happy to discuss further and clarify any extra item 🙂