Legacy Modernisation, No Downtime
Old and new running side by side until the new path has earned the traffic. Nothing is switched on a single weekend: routes move a slice at a time, differences are investigated rather than waved through, and the previous path stays available the whole way through.
- To first release
- 2 weeks
- Downtime
- 0
- Kafka
- Go
- Kubernetes
- Terraform
- Postgres
The problem
A revenue-critical monolith is expensive to change and risky to touch, but a rewrite-and-switch cutover puts the whole business on a single deploy. Neither standing still nor a big-bang rewrite is acceptable.
How we would sequence it
- 01Map real traffic and data flows before touching any code
- 02Put an event bridge between the monolith and the new services
- 03Route a small traffic slice to the new path and compare outputs
- 04Widen the slice only when parity holds under production load
- 05Keep the rollback switch live until the old path is decommissioned
This solution describes our method for a class of problem. It is not a record of a completed client engagement, and the figures above are scope estimates rather than measured results.
Have this problem?
Scope this with us