feat(quota): add usage tracking for Groq

First slice of #3701: quota tracking for Groq via x-ratelimit-* response
headers on the models endpoint (no dedicated quota endpoint exists, and
reading usage costs zero tokens).

- usage/groq.js: parse request+token limit/remaining headers; Go-style
  duration reset headers ("2m59.56s") resolve to future timestamps;
  missing key/401/403 -> message, 2xx without headers -> soft
  "not tracked yet" with quotas:{}
- registry/groq.js: transport.usage.url (reuses validateUrl) +
  features {usage, usageApikey}
- services/usage.js: groq entry in USAGE_HANDLERS
- ProviderLimits/utils.js: parseQuotaData case (absolute used/total,
  codex/kiro style)
- tests: groq-usage.test.js (registry flags, header parsing, soft
  not-tracked path, missing key/401, parseQuotaData)

open-sse/providers/registry/groq.js

@@ -17,6 +17,12 @@ export default {
   transport: {
     baseUrl: "https://api.groq.com/openai/v1/chat/completions",
     validateUrl: "https://api.groq.com/openai/v1/models",
+    // No dedicated quota endpoint; rate-limit info rides on x-ratelimit-*
+    // response headers, always included. Reuse the models list (already
+    // used as validateUrl) so reading usage never costs tokens.
+    usage: {
+      url: "https://api.groq.com/openai/v1/models",
+    },
   },
   models: [
     { id: "llama-3.3-70b-versatile", name: "Llama 3.3 70B" },
@@ -34,4 +40,8 @@ export default {
     authHeader: "bearer",
     format: "openai",
   },
+  features: {
+    usage: true,
+    usageApikey: true,
+  },
 };