burla

Provision and use isolated ephemeral GCP VMs for Burla local-dev or remote-dev. Use when the user mentions Burla dev VMs, local-dev or remote-dev on GCP, SSH tunnel dashboards, isolated agent runtimes, or ephemeral VM workflows.

Detailed reference for Burla subsystems - end-to-end job lifecycle for remote_parallel_map, cluster and node lifecycle (boot, grow, reserve, shutdown), and Firestore schema. Use when working on Burla internals, debugging cross-service behavior in the burla repo, tracing how a job flows between client / main_service / node_service / worker_server, or answering detailed questions about how a specific Burla subsystem works.