fix(metering): register the gateway's api root, not the gateway root

[pyramation]

Summary

A metered model registered the gateway root as its provider baseUrl, so every real completion went to POST <root>/chat/completions and the gateway answered 404 Cannot POST /chat/completions — it serves the OpenAI-compatible surface under /v1. Found by running a coding-agent task in-cluster against agentic-server, where a model turn died instantly with that 404.

resolveMeteredGateway now registers the api root, and the root normalizer tolerates a caller that already spelled /v1 so a configured https://gw.example.com/v1 cannot become /v1/v1:

normalizeGatewayUrl('https://gw.example.com/v1') // 'https://gw.example.com'
completionsBaseUrl('https://gw.example.com')     // 'https://gw.example.com/v1'
resolveMeteredGateway({ gatewayUrl, ... }).baseUrl // <root>/v1

Usage reporting is unchanged (<root>/v1/usage), and completionsBaseUrl is exported so a harness that builds its own provider config asks for the same route rather than re-deriving it.

Link to Devin session: https://app.devin.ai/sessions/2d292e43fd2e4f34bb84605b5aa2a250
Requested by: @pyramation

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment, CI, and merge conflict monitoring