fix(gateway): track uptime for used provider

[steebchen]

Summary

• Fixes uptime tracking to use the actual provider that handled a request, not the originally requested one
• When low-uptime fallback routes to a different provider, stats are now correctly attributed to the fallback provider
• Fixes test data inconsistency where usedModel had wrong provider prefix

Test plan

• Existing stats-calculator tests pass (17 tests)
• Provider-metrics tests pass (19 tests)
• Build succeeds
• Verify uptime stats are correctly attributed after low-uptime fallback in production

🤖 Generated with Claude Code

Summary by CodeRabbit

• Bug Fixes

  • Improved model identifier resolution to ensure the correct model is used for statistics and logging when routing or fallback occurs.

• Tests

  • Aligned test data to ensure model and provider pairs are correctly matched in statistics calculations.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[coderabbitai]

Walkthrough

Changes the baseModelName selection logic in the chat handler to prefer the canonical model ID from modelInfo when available, with a fallback order of modelInfo.id, finalModelInfo?.id, then usedModel. Updates corresponding test data to ensure model/provider pairs align properly.

Changes

Cohort / File(s)	Summary
Model ID Resolution apps/gateway/src/chat/chat.ts	Modified baseModelName calculation to prioritize canonical ID from modelInfo (when present as ModelDefinition) before falling back to finalModelInfo?.id and usedModel. Ensures correct model identifier is used for routing/fallback logic and downstream stats/logging.
Test Data Alignment apps/worker/src/services/stats-calculator.spec.ts	Updated test entry to align usedModel (anthropic/gpt-4) with usedProvider for GPT-4 mapping, ensuring model/provider pairs match in statistics calculations.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Possibly related PRs

• PR #1230 — Directly related: modifies routing/fallback behavior that updates usedProvider/usedModel and routingMetadata in the same file.
• PR #1291 — Related: changes model resolution behavior in apps/gateway/src/chat/chat.ts to refine model lookup logic.
• PR #675 — Related: modifies how the base model identifier is determined and propagated through routing/fallback paths.

Suggested labels

auto-merge

Suggested reviewers

• smakosh

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title 'fix(gateway): track uptime for used provider' accurately summarizes the main change: correcting uptime tracking to attribute stats to the actual provider that handled requests rather than the originally requested provider.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}