User Guide: Running Pulumi in Production

[bermudezmt]

• How does one run Pulumi-defined infrastructure in production? We support Azure DevOps, which is good. More effort here is needed.

• Give some further guidance on how to use stacks vs. environments and resource groups, how the state works, what might be some situations when you need to interfere with the state.

• How to run blue-green deployments, what would a typical CI/CD pipeline look like? Testing in production, monitoring, alerts, dashboards, and how Pulumi helps with that (if it does). Maybe another reference solution that includes all that.

[joeduffy]

Also see https://livebook.manning.com/book/terraform-in-action/chapter-9/v-4/15.

[joeduffy]

A bit of user feedback right now that would be good to address with this effort:

In addition to stacks, state backends and where to store your programs is a little unclear. It feels like some knowledge is assumed here. For instance, if I choose an S3 bucket to store my state information, what happens is a .pulumi directory is created with state history for my project. The thing that's unclear is where do I store my project? Is it assumed my project/program should be stored in git some where? This part of the solution doesn't appear to be documented anywhere and no suggestions are provided on where to store stack state, project state and program state.

[joeduffy]

Example user guide we could add here:

Pulumi in Production

• Organizing Stacks
• Structuring Individual Projects
• Structuring Multi-Project Environments
• Adopting Continuous Delivery
• Adopting Continuous Verification

[cnunciato]

Note that we've since added the Zephyr project, which covers a good bit of this surface area. https://www.pulumi.com/blog/iac-recommended-practices-code-organization-and-stacks/