We put up a post today on how we use Pulumi to build Pulumi. A few interesting examples of use cases we've found ourselves for Pulumi, and pointers to some of the libraries we've built to make similar use cases easier for others to build on top of. http://blog.pulumi.com/how-we-use-pulumi-to-build-pulumi

awesome post! curious about strategies (more detail) for working with the standalone eng/dev instances

Right now, many folks have their own long-lived instances of the backend infrastructure, and iterate on them from their local dev environment. Others who touch the service code less regularly bring up a dev stack only when needed. Since we are deploying containerized services, developers can easily run

update

locally on their stack, which will deploy infrastructure updates as well as build and push container images from their local environment. This provides a pretty fast inner loop for deploying both infrastructure changes and application code into the dev environment. A similar process applies to out more permanent testing, staging and production environments - but these are deployed via CI/CD, not from local developer machines. CC also @square-apartment-28429 and @powerful-whale-8767 who may have more details to share here. We'll also be discussing this live on our YouTube stream tomorrow with Chris and Matt, so feel free to join us and/or share any additional questions you'd be interested in us going into more detail on.

right on thanks for the quick and thorough reply!