There are for example https://spacelift.io/ci-cd-for-infrastructure And this, that claims to do it all using AI https://circleci.com/
From: Silklist <[email protected]> on behalf of Manar Hussain via Silklist <[email protected]> Date: Tuesday, 28 October 2025 at 4:37 PM To: Alaric Snell-Pym <[email protected]>, Intelligent conversation <[email protected]> Cc: Manar Hussain <[email protected]> Subject: Re: [Silk] A digital Plan B On 28/10/2025 18:37, Alaric Snell-Pym wrote: > On 27/10/2025 14:41, Manar Hussain wrote: > >> So what's the next barrier? My take: deployment -- the effort to >> choose/ assess, get buy-in/communication, configure, deploy, maintain >> and extend etc. > > Yes! > > I was recently feeling frustrated about the complexity of "CI" setups > I've encountered at different clients. All of them have some kind of > gitlab or github pipeline set up to perform linting of the code and > tests, and ways to run those same lints and tests in your working copy > before committing, and usually some infrastructure in a cloud provider > of choice to provide a test environment where things can be deployed > to from CI or locally to run the tests against them. And often a > docker image that's set up to run the non-cloud-hosted parts of the > tests from, such as the linters and "terraform apply" and to run tests > that submit sample inputs to the stuff terraform has set up and check > the results are correct. > > It's quite a lot of stuff. Sometimes thousands of lines of scripts and > configuration files. Yet every project has to build it themselves, > more or less from scratch. And then I come and join the project and > need to learn the particular spellings of everything each project > uses, and all its particular quirks and complications, and where the > scripts are all kept, and and and... It just seems a lot of > duplication of effort to create very similar results each time! I spoke with a founder of a really interesting UK company that supplies devops engineers for various uk gov departnes, as well as training AND trains devops engineers from scratch -- all knitted together to help departments get their in house team scaled up and working in a coherent way (i.e. with broadly the same norms/approach), with the gap bridged during the transition bridged by per day contractors they make good money on. I.e. they are already "selling" some level of standardisations because that's really useful even though there are many variants on what would make an OK standard. And yet even this company was intrigued but unwilling to invest in deployment as a product because they get paid per diem for all the effort it takes to reinvent the wheel so why should they. It'll come, especially with AI maturing, but I've been waiting a long time for it. -- Silklist mailing list [email protected] https://mailman.panix.com/listinfo.cgi/silklist
-- Silklist mailing list [email protected] https://mailman.panix.com/listinfo.cgi/silklist
