I've recently become a fan of more dynamic IaC like the AWS CDK or Pulumi, since I can write more sophisticated stack deployment logic using my skills as a TypeScript developer and sometimes share common code between application deployment logic and infra deployment logic, etc.
I can agree that solving deployments isn't where architectural problems end, but without seeing this IaDE I can't really wrap my head around it, and wherever I navigate is just a placeholder, and I see nothing on GitHub. Is this also code-based? Does this integrate monitoring more easily into stack deployments?
I’m with you on CDK or Pulumi. Being able to use programming languages for infra makes composition better.But the main issue is that they still rely on the same execution model: define intent, generate a plan, apply, and then the process exits. The system doesn’t stay "alive" after that. If something drifts, partially fails, or never fully converges, you’re back to fixing it manually.
With Planetform, the idea is a shift to Durable Execution (IaDE). It’s less about how we define infra and more about making the system itself responsible for staying correct over time. Monitoring then becomes part of that verification loop, not something you wire in after the fact.
This is the first of a couple of posts we’ll be releasing on how we think about solving this. The project will be fully open source
I've recently become a fan of more dynamic IaC like the AWS CDK or Pulumi, since I can write more sophisticated stack deployment logic using my skills as a TypeScript developer and sometimes share common code between application deployment logic and infra deployment logic, etc.
I can agree that solving deployments isn't where architectural problems end, but without seeing this IaDE I can't really wrap my head around it, and wherever I navigate is just a placeholder, and I see nothing on GitHub. Is this also code-based? Does this integrate monitoring more easily into stack deployments?
I’m with you on CDK or Pulumi. Being able to use programming languages for infra makes composition better.But the main issue is that they still rely on the same execution model: define intent, generate a plan, apply, and then the process exits. The system doesn’t stay "alive" after that. If something drifts, partially fails, or never fully converges, you’re back to fixing it manually. With Planetform, the idea is a shift to Durable Execution (IaDE). It’s less about how we define infra and more about making the system itself responsible for staying correct over time. Monitoring then becomes part of that verification loop, not something you wire in after the fact. This is the first of a couple of posts we’ll be releasing on how we think about solving this. The project will be fully open source