persist studio container

[themightychris]

Create a projection CLI option/env var that is enabled by default that sets a max idle time for studio containers to e.g. 30 minutes.

The studio container would self-assess when its last command finished and kill itself once that timeout elapses. Future projection commands on the host would look for an existing container and re-use if running instead of starting a new one. Studios could be re-used across any command and in parallel so no locking or unique identification should be necessary—any running container that is available should be usable.

[themightychris]

Lens workspace reuse could be a problem with parallel execution—maybe they need to be re-architected to use uniquely-identified temp directories instead anyway