Skip to main content

Welcome!

Why Salesforce DevOps Needs Unique Treatment is the first course in the Flosum Certification. From first principles, it makes the architectural case for why standard Git-based DevOps breaks on Salesforce — and what kind of system actually fits.

For: Developers, architects, DevOps engineers, technical leads, & platform owners

Length: 8–10 hours, over 2–3 weeks

Prerequisites: Knowledge of Git, CI/CD fundamentals, Salesforce Metadata API

Key takeaways​

By the end of this course, you will be able to:

  1. Explain why Salesforce's metadata model breaks Git's content-addressed assumptions, with reference to the Metadata API.
  2. Name the seven mismatches by mechanism: what fails, why it fails, and what kind of system would prevent it.
  3. Build a defensible cost model that maps these failures to DORA metrics and to your team's actual deployment data.
  4. Ask the one diagnostic question that separates a Git-wrapper tool from a metadata-native system.
  5. Specify, at component level, what a metadata-native architecture must do, and where the two architectural schools (extend Git vs. replace it) actually differ.
  6. Diagnose where your own pipeline sits relative to DORA Elite, and identify which mismatches are costing you the most.

Lessons​

Salesforce is categorically different from the software Git was built for, and the difference is architectural, not cosmetic. That mismatch produces seven specific failure modes — each one rooted in something Git was never designed to do. The fix isn't a better wrapper around Git. It's a system that reasons about Salesforce metadata semantically, from the deployment engine up.

Each lesson takes one step of that argument and grounds it in a technical mechanism, a failure scenario, and the root cause that links the two.

1. Getting started

Dive into DevOps and Salesforce, and understand the pillars of this course.

2. Two Different Worlds

See Salesforce vs. Git from the Metadata API and content-addressed vs. semantically addressed systems.

3. The Seven Mismatches

Examine each mismatch by technical mechanism, failure scenario, and root cause.

4. The Economics of Drift

Build the formal cost model, map it to DORA metrics, and work through the sensitivity ranges.

5. Why Patching Git Doesn't Work

Learn the diagnostic question that separates wrapper tools from metadata-native systems.

6. The Architecture You Actually Need

Specify the technical components of a metadata-native system, and where the two architectural schools differ.