docs: add ClawMetry session observability tutorial

[vivekchand]

Closes #11282.

Adds a tutorial page for observing goose sessions with ClawMetry, following the structure of the existing Langfuse and MLflow observability pages.

@jamadeo you asked whether this works without a paid plan. It does now. goose is a free runtime in the open source app, and the adapter ships in the package under MIT, so the whole path is self-hosted:

• clawmetry/adapters/goose.py
• FREE_RUNTIMES in the entitlement resolver lists goose alongside openclaw and nemoclaw

What the page covers. pip install clawmetry, open http://localhost:8900, and existing goose sessions appear. There is no goose-side configuration step because ClawMetry is not in the request path: it reads $XDG_DATA_HOME/goose/sessions/sessions.db read-only, honouring XDG the same way goose does.

On accuracy. The page claims only what the adapter actually surfaces, which is sessions, transcripts, models, and token totals. It states plainly that goose records a cost figure only for providers that report one, so with a local provider such as Ollama the tokens are real and the cost is left empty rather than estimated. No live-stream or agent-control claims, since the goose adapter does not declare those capabilities.

Notes for review.

• The tutorials sidebar is autogenerated, so no sidebars.ts change is needed.
• No screenshot is included yet. The dashboard renders transcript content, so I would rather shoot it against a session I can safely publish than paste one in. Happy to add an image in a follow-up commit if you want one for parity with the other pages.

I am the ClawMetry maintainer. Glad to adjust tone, length, or structure to whatever you prefer.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 57f34881a0

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]

Use the platform session-store path

For default macOS installs, this is not where goose writes the sessions DB: Paths::data_dir() uses the platform app strategy, with the repo’s own env-var guide listing macOS as ~/Library/Application Support/Block/goose/, and SessionStorage then appends sessions/sessions.db. As written, macOS users following the “no configuration” flow are told ClawMetry will auto-detect ~/.local/share/..., so their existing goose sessions will not appear unless the tutorial or integration accounts for the actual macOS path.

Useful? React with 👍 / 👎.