# getting-started


11/30/2023, 9:52 AM
Hey everyone, quick "best practices" question. We have as of now a gigantic graph of all our resources. Meaning one stack is deploying everything from the load-balancer to every application within the cluster, comprised with everything needed (dns, iam, buckets etc ...). Our main issue is that one stack is a bottleneck when a deployment is failing, and it's getting difficult to work as a team on it, plus it's getting slower as we add more and more resources. What's the good way to handle a large infrastructure with pulumi. Should we have one stack per "app", should we do something else? I'm starting to enter an endless debate with myself and would love to have some insights. Thanks!


11/30/2023, 2:53 PM
@millions-iron-7317 may I refer you to our blog series on IaC Best Practices as a start on giving you new ideas?
Do read all articles first, as later articles revise the setup while the (fictitious) Zephyr company is growing their infrastructure.