BlueGreen
Two full, identical environments (blue = current, green = new); traffic cuts over to green all at once, and blue stays live as an instant rollback target.
Why we need this / what value this brings
Rollback becomes a traffic-routing decision (seconds) instead of a redeploy (minutes to hours) — the old version is still fully running and ready.
When to use this
When rollback speed matters more than the cost of running two full environments simultaneously, even briefly.
How to use or implement this
Deploy the new version to a fully separate environment, run PIV/smoke tests against it before cutting traffic over, and keep the old environment warm for a defined window after cutover in case a rollback is needed.
Research questions
- Instant rollback (just cut traffic back to blue) is the big advantage — why wasn't an equivalent 'keep the old golden image live and just switch back' available in the PIV-failure scenario?
Empty folder — drop notes, links, and findings here as you research.