build: collector service Dockerfile

[Svaag]

Adds the container image for the trace collector so the loop host can docker build it
like knowledge-mcp.

• Dockerfile — python:3.12-slim + uv, installs agent-core[collector], runs
  python -m agent_core.collector (FastAPI); HYRULE_COLLECTOR_DATABASE_URL selects the DB
  (postgresql+asyncpg in prod).
• .dockerignore.

Runtime validated by the collector TestClient suite (#4). No docker daemon in CI/dev here, so
the image build is validated on the target host by the ansible collector role (C3).

🤖 Generated with Claude Code

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 709131bcec

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Configure a usable default database

When this image is launched without HYRULE_COLLECTOR_DATABASE_URL (the Dockerfile does not set one), agent_core.collector.db falls back to sqlite+aiosqlite:///./collector.db. That default is not usable in this image: the install here only pulls agent-core[collector], which omits aiosqlite, and after switching to UID 65534 the process also cannot create collector.db under the root-owned /app. A plain docker run therefore exits before /healthz is served unless every caller supplies a Postgres URL; either make the sqlite fallback work in the image or fail/configure explicitly around the required database URL.

Useful? React with 👍 / 👎.