[refactor] use new api container lifecycle bench

[mathewjhan]

What

Why

How to Test

Checklist

• PR title follows [module] type: description format
  (labels are derived from it automatically — no need to add them by hand)
• ruff check . and ruff format --check . pass
• pyright osmosis_ai/ passes
• pytest passes (new tests added if applicable)
• Public API changes are documented
• No secrets or credentials included

---

Summary by cubic

Refactors the container lifecycle benchmark to use the RolloutClient API instead of a custom controller server.

• Removes the in-benchmark controller app that received completion callbacks.
• Submits rollouts through RolloutClient.run_rollout_async and waits on the returned future.
• Keeps the same timeout behavior by doubling the per-rollout timeout to cover both agent and grader.

^{Written for commit 48aee7b. Summary will update on new commits.}