Cloud and Platform Engineering Foundations

Cloud engineering is not a console. It is a way of working: you describe the state you want in code, you review the change before it happens, you keep the description and the reality in step, and you run the result for other people. This course teaches that way of working on a machine you can see all of, with no vendor subscription anywhere. Your lab machine **plat01** is your platform. Its own Docker engine is the cloud you provision: OpenTofu creates the networks, volumes and containers, and you can `docker inspect` every resource you declared. Over ten lessons you build one repository for a synthetic internal team - Team Kestrel at Fennimore Diagnostics - and it grows the way a real platform repository grows: - the vocabulary of cloud computing in vendor-neutral terms, mapped to primitives you measure on your own machine, and what a platform team actually provides; - OpenTofu from zero: HCL, providers, `init`, `plan`, `apply`, state, the lock file, `destroy`; - modules, input validation, `moved`, drift you cause on purpose and then detect and reconcile, and adopting a resource somebody made by hand with `import`; - a two-colour web service with a reverse proxy on a user-defined network, with your OpenTofu outputs feeding an Ansible inventory, continuing the work you did in ULC-106; - machine images, cloud-init user data you validate and build into a NoCloud seed, and a Proxmox VE virtual machine written and validated - never applied, because this lab has no hypervisor and the course says so; - a Git pipeline whose `pre-receive` gate really refuses a badly formatted, invalid or secret-carrying push; - secrets with age and sops, OpenTofu state encryption, and a secret scanner that finds a real leak; - health checks, a probe timer writing to journald, a backup you restore into a new container and verify, and a zero-downtime image replacement measured with a request loop. Two lessons start from a deliberately seeded fault - a credential a colleague pasted into a variables file, and a service somebody "fixed" by hand on a Friday - and you work them the way an operations team does: observe, form one hypothesis, test one thing, fix, verify, record. The final project is a second environment for the same team, delivered with a change record, a runbook, a restore drill and an architecture note. This is a learning pathway toward platform engineer and cloud administrator roles. It awards an Ultiblob Certificate of Completion; it is not a vendor certification and makes no promise about employment.