An AI agent writes code, and now you have to decide where it runs, because in your process a single prompt injection reaches everything. Three answers sit at three points on the one axis that matters, isolation strength - plain Docker (thin shared-kernel wall), my tank-os-with-llms container (that thin wall hardened, self-hosted), and Azure Container Apps Sandboxes (thick hardware-backed microVM, cloud-managed). Isolation strength and ownership pull against each other, and here is how to pick.
tag: #containers
2 posts
2 weeks ago
Where Should Agent Code Run
PID 1 in Linux has special responsibilities that most application processes are not designed to handle. The --init flag in Docker inserts a tiny init process between PID 1 and your application, which solves the zombie reaping problem and the signal forwarding problem in one line.