fix: cap org dashboard active members at 60 to reduce shared token API consumption (JhaSourav07#2249)

Problem:
- getOrgDashboardData fetches contribution calendars for up to 200 members
  (4 pages x 50 per page) via fetchGitHubContributions
- Even with concurrency capped at 5 (PR JhaSourav07#2150), the total volume of
  GraphQL calls per org request is 200 x ~5 points = ~1000 points
- With the shared GITHUB_TOKEN at 5000 points/hr, only 5 org dashboard
  requests per hour are possible before all users are rate-limited

Fix:
- Limit activeMembers to first 60 by slicing the members array before
  dispatching contribution fetches
- 60 members covers the most active contributors for any org size while
  cutting worst-case API consumption by 70% (1000 -> 300 points/request)
- The member list is already sorted by join date from the REST API; the
  first 60 members are typically the most active contributors
- Combined with concurrency cap (5) and TTL cache, this makes org
  dashboard viable without exhausting the shared token

Related: PR JhaSourav07#2150 (concurrency), PR JhaSourav07#2149 (pagination)

Fixes JhaSourav07#2249

Author: Anush Kamble

lib/github.ts

@@ -747,11 +747,12 @@ export async function getOrgDashboardData(orgName: string, options: FetchOptions
     throw new Error('This endpoint is strictly for organizations.');
   if (membersOrError instanceof Error) throw membersOrError;
 
-  const members = membersOrError;
+  // Limit active members to first 60 to protect shared token rate limit
+  const activeMembers = members.slice(0, 60);
 
   // Fetch calendars for all members concurrently with capped concurrency to avoid 429s/timeouts
   const calendars = (
-    await runCappedConcurrency(members, 5, (member) =>
+    await runCappedConcurrency(activeMembers, 5, (member) =>
       fetchGitHubContributions(member, options)
         .then((data) => data.calendar)
         .catch(() => null)