How environments drift apart
- A personalisation or script is changed in production to solve an urgent issue, and never brought back to test.
- A data flow or interface is set up in acceptance for a project, and forgotten.
- A setting is changed in one environment during a test and not reset.
- A refresh brings data over, but not the configuration around it, or the other way round.
What to compare
| Area | What to look for |
|---|---|
| H5 personalisations and scripts | Present in one environment only, or different in content |
| ION data flows | Flows that exist, or are active, in one environment only |
| MEC agreements | Differences in mapping or partner set-up |
| Interfaces | Endpoints and connections that differ between environments |
| Key settings | Values that steer processes, such as order type or warehouse settings |
Compare snapshots, not memories
The most reliable comparison is between documented snapshots of each environment, taken from the environment itself. They show what differs now, and also what changed since the last snapshot: often the fastest way to find the cause when something suddenly behaves differently.
Compare the test environment with production first. A known difference is a decision; an unknown difference is a false test result waiting to happen.
Make it routine
- 1Take regular snapshots of every environment.
- 2Review the differences between PRD and the test environments before each test cycle and each release.
- 3Decide per difference: bring it in line, or document why it is different.
- 4Keep the documentation generated from the environments themselves, so it never goes out of date.