fix(cache): prevent unbounded TTLCache memory exhaustion and runtime instability (JhaSourav07#548)

## Description

Fixes JhaSourav07#537

This PR implements a bounded-capacity eviction strategy for the global
`TTLCache` instances used inside `lib/github.ts`.

Previously, the cache relied only on TTL expiration and did not enforce
any maximum capacity limits. Under sustained unique-user traffic
(crawler indexing, bots, traffic spikes, or large username scans), the
internal `Map` continuously retained strong references for new cache
entries, causing unbounded heap growth and increasing the risk of
Node.js/Vercel runtime instability or OOM crashes.

This fix introduces explicit cache ceilings while preserving the
existing TTL lifecycle and maintaining full backward compatibility.

### Changes Implemented

- Added bounded cache size limits for:
  - `contributionsCache`
  - `profileCache`
  - `reposCache`

- Added production-focused defensive comments explaining:
  - memory exhaustion risks
  - crawler amplification scenarios
  - runtime stability concerns
  - bounded eviction rationale

### Root Cause

The cache instances were initialized without supplying `maxSize`:

    const contributionsCache = new TTLCache<ContributionCalendar>();
    const profileCache = new TTLCache<GitHubUserProfile>();
    const reposCache = new TTLCache<GitHubRepo[]>();

Because the internal `Map` retained strong references to active entries:
- memory scaled linearly with unique usernames
- garbage collection could not reclaim active objects
- long-lived processes accumulated memory continuously

### Validation

Successfully verified with:

- `npm run test`
- `npm run lint`
- `npm run build`

### Test Results

- 22/22 test suites passed
- 275/275 tests passed
- Production build compiled successfully

### Runtime Validation

Stress-tested with 50,000 unique usernames.

| Metric | Result |
|---|---|
| Heap Before | 7.41 MB |
| Heap After | 15.55 MB |
| Behavior Before Fix | Unbounded growth |
| Behavior After Fix | Stable bounded plateau |

### Impact

This PR prevents:
- unbounded heap growth
- excessive GC pressure
- potential Vercel OOM crashes
- runtime instability under crawler/bot traffic

while preserving:
- existing TTL behavior
- cache semantics
- API compatibility
- backward compatibility

---

## Pillar

- [ ] 🎨 Pillar 1 — New Theme Design
- [ ] 📐 Pillar 2 — Geometric SVG Improvement
- [ ] 🕐 Pillar 3 — Timezone Logic Optimization
- [x] 🛠️ Other (Bug fix, refactoring, docs)

---

## Visual Preview

N/A — backend runtime stability fix

---

## Checklist before requesting a review:

- [x] I have read the `CONTRIBUTING.md` file.
- [x] I have tested these changes locally
(`localhost:3000/api/streak?user=YOUR_USERNAME`).
- [x] I have run `npm run format` and `npm run lint` locally and
resolved all errors.
- [x] My commits follow the Conventional Commits format.
- [x] I have started the repo.
- [x] I have made sure that I have only one commit in this PR.
- [x] All tests and production builds pass successfully.

Author: JhaSourav07

lib/github.ts

@@ -114,9 +114,20 @@ interface GitHubUserProfile {
   plan?: { name?: string } | null;
 }
 
-const contributionsCache = new TTLCache<ContributionCalendar>();
-const profileCache = new TTLCache<GitHubUserProfile>();
-const reposCache = new TTLCache<GitHubRepo[]>();
+// Named constants to avoid magic numbers and allow future tuning
+const MAX_CONTRIBUTIONS_CACHE_SIZE = 1000;
+const MAX_PROFILE_CACHE_SIZE = 1000;
+const MAX_REPOS_CACHE_SIZE = 500;
+
+// Bounded capacity controls to prevent unbounded heap memory leaks (OOM).
+// Under continuous crawler/bot scanning or viral peaks, unbounded cache size
+// allocations will exhaust Node/Vercel serverless RAM.
+// Specifying explicit capacity limits enforces a First-In, First-Out (FIFO)
+// eviction strategy (since standard ES6 Map maintains key insertion order) and
+// bounds max memory consumption to stable, predictable boundaries.
+const contributionsCache = new TTLCache<ContributionCalendar>(MAX_CONTRIBUTIONS_CACHE_SIZE);
+const profileCache = new TTLCache<GitHubUserProfile>(MAX_PROFILE_CACHE_SIZE);
+const reposCache = new TTLCache<GitHubRepo[]>(MAX_REPOS_CACHE_SIZE);
 
 function cacheKey(
   kind: 'contributions' | 'profile' | 'repos',