Narrow Containerfile build targets to tracker image needs

[josecelano]

Goal

Reduce container image build time by avoiding compilation and linking of workspace targets that are not required to produce and validate the tracker runtime image.

Background

The current Containerfile builds and archives a very broad target set (--tests --benches --examples --workspace --all-targets --all-features) across multiple stages. A quick maintainer analysis suggests some of that work is unrelated to the final tracker image, including targets from packages such as packages/torrent-repository-benchmarking.

This issue should only proceed after the baseline subissue confirms both of these points:

1. Unneeded target compilation/linking is materially present in the container build path.
2. That work has significant impact on workflow runtime.

If confirmed, narrowing target scope can speed up container.yaml directly, and can also improve testing.yaml because Docker E2E builds and uses the tracker image there.

Acceptance Criteria

• AC1: Baseline evidence confirms that unnecessary target compilation/linking is a significant bottleneck.
• AC2: Containerfile target scope is reduced without removing artifacts required by the runtime image.
• AC3: Container workflow runtime improves measurably after the change.
• AC4: Testing workflow Docker E2E path remains valid and does not regress.

Related

• Parent EPIC: EPIC: Improve PR Workflow Performance #1840
• Baseline analysis: Baseline workflow profiling and bottleneck analysis (EPIC #1840) #1841
• Full spec: docs/issues/open/ (after spec PR merge)