You've asked the quintessential "where do I start" question, which usually invites a flood of marketing blog posts and vendor checklists that are worse than useless. They'll have you comparing superficial feature grids without understanding the foundational costs—both in engineering hours and cloud spend—that these platforms lock you into.
As a self-proclaimed complete newbie, your first step isn't to compare tools. It's to instrument your own manual build and deployment process so you understand what you're actually automating. You cannot evaluate a CI/CD platform if you don't know what your pipeline *does*. So, before you look at a single tool's website, do this:
1. Pick your simplest application.
2. Write down, in painful, granular detail, every single command you run from `git clone` to a running service. This includes dependency installation, compilation, unit tests, building a container, pushing it, updating a Kubernetes manifest or Lambda function, and any post-deployment smoke checks.
3. Time each step. Note where it fails. Document the exact compute environment (your laptop specs, probably).
Now you have a benchmark. A terrible, manual, non-scalable benchmark, but a real one. Convert that list into the most basic, ugly shell script possible. Make it run on a spare machine or a cheap cloud VM. That shell script is your baseline CI/CD system. Its runtime and failure points are your first performance metrics.
Only *now* should you start looking at tools. Your comparison criteria should shift dramatically from "does it have a pretty UI?" to concrete questions like:
* How does this platform execute my exact shell script steps? Do I have to rewrite them into their proprietary YAML dialect, or can I run the script largely as-is?
* What is the *real* cost model? If they promise "free for open source," what are the concurrency limits? If it's SaaS, what's the egress cost for moving multi-gigabyte artifacts to my own cloud? If it's self-hosted, what's the operational overhead of maintaining the runners?
* Where does state live? How are secrets handled? Can I easily replicate the entire pipeline configuration in another tool, or am I buying into a high-walled garden?
You'll see enthusiastic reviews praising "developer experience" and "integrations." Ignore them. Look for the complaints. Search for "[Tool Name] outage," "[Tool Name] cost surprise," and "[Tool Name] migration horror story." The negative experiences tell you far more about the platform's architecture and business priorities than any shiny feature list.
So, start with your shell script. Post its steps and its pain points here, and we can give you a brutally honest comparison of which platforms will exacerbate those pains and which might actually alleviate them. Without that concrete baseline, you're just shopping for brand names.
Trust but verify.