Collapsing Lambda into Containers
A multi-agent operations platform that audits, hardens, prioritizes, and tracks work across the rest of the portfolio. Back Office is built around transparent automation, operator control, and privacy-first hardening instead of surveillance-heavy workflow tooling.
Common questions
Why not keep Lambda if it was already working?
It was working, but the shape was wrong. Every route meant another function, another packaging step, another cold start profile, and another piece of API Gateway config. The total surface area outgrew what the workloads actually needed.
What is a Magic Container on Bunny?
A container runtime on Bunny's edge infrastructure. You give it a Docker image, it runs it close to your users. It replaces the Lambda plus API Gateway combination with a single deployable unit that handles its own routing.
Did you lose anything by leaving Lambda?
Per-function scaling granularity. On Lambda each function scales independently. In a container you manage that yourself. For this portfolio the tradeoff was easy because no single route had dramatically different scaling needs from the others.
How did the deployment story change?
It got shorter. Instead of packaging individual functions, wiring API Gateway routes, and deploying through SAM or CDK, I push a container image and Bunny runs it. The deploy script went from orchestrating multiple AWS services to one image push.