I used this view on Looker to determine what sort of deployment the interested customers are using. Found it was mostly kubernetes, but also some docker compose, pure docker, and docker container deployments. I think it's important to understand where this tool needs to be run, and what kinds of differences there might be between deployments.
My plan was to use the debug tools (debug-server, debug-compose, debug-kube) to help serve as a template as to how to interact with the different types of deployments. I started with single docker since it was easy to run locally and seemed the most straight-forward to start with.
I didn't get very far with the implementation, and ran into an issue where it looks like pg_dump was not being found as a command. I was investigating how to fix that, but had not yet found a solution.
There is also the piece of uploading the output to the cloud, which I had not started on.
Test plan
I was planning to do some similar testing as done in this PR to do initial testing on a variety of environments. I think other useful environments to test on could be k8s and s2.
