[PERF] Cache GitHub language API responses — currently refetches on every dashboard load

[Priyanshu-byte-coder]

Problem Statement

Language data from GitHub's /repos/{owner}/{repo}/languages endpoint is fetched fresh on every dashboard load. Language composition of a repo changes rarely — daily at most. This wastes GitHub API rate limit quota and slows down the repos widget.

Proposed Solution

Apply the existing withMetricsCache wrapper to the languages fetch with a long TTL (e.g. 6 hours). Language data for a given repo almost never changes intraday.

Alternatives Considered

• No cache: Current behavior — wastes rate limit
• Infinite cache: Too aggressive — user won't see new language additions

Acceptance Criteria

• Languages fetch wrapped with withMetricsCache and TTL = 21600 seconds (6h)
• Cache key includes repoName to avoid cross-repo collisions
• Cache is bypassed if user clicks Refresh on the widget

Additional Context

The METRICS_CACHE_TTL_SECONDS object in src/lib/metrics-cache.ts should get a languages: 21600 entry. This follows the same pattern as all other metric caches.

[harshitanagpal05]

pls assign this issue to me?

[mallya-m]

Hi @Priyanshu-byte-coder!

I've already contributed to this repo with 5 merged PRs (#303, #411, #586, #602, #306) and I'd love to fix this performance issue.

My approach:

• Add languages: 21600 to METRICS_CACHE_TTL_SECONDS in src/lib/metrics-cache.ts
• Wrap the languages fetch in src/app/api/metrics/repos/route.ts with withMetricsCache
• Cache key includes repoName to avoid cross-repo collisions
• Cache bypassed when user clicks Refresh (existing bypass pattern already in codebase)

I'm already familiar with withMetricsCache from my previous contributions to this repo — this is a clean 2-file change.

Could you please assign this to me? GitHub: github.com/mallya-m

[tmdeveloper007]

Proposed Implementation Plan

Here is a detailed, technically solid, step-by-step approach to solve this issue:

1. Open the languages metric endpoint (e.g. in /api/metrics/repos or dedicated route).
2. Wrap requests with the withMetricsCache middleware wrapper, and set a TTL entry languages: 21600 (6 hours) in src/lib/metrics-cache.ts.
3. Append the repository name to the cache keys to isolate entries, and bypass checks if the user sends a refresh query flag.

I would love to work on this issue and implement the proposed solution. Could you please assign this issue to me? Thank you!

[Sammmyyyyyyy]

Hi @Priyanshu-byte-coder!

I'd like to work on this issue under GSSoC'26.

Approach:

• Add caching for the GitHub languages fetch using the existing withMetricsCache utility
• Configure a dedicated languages TTL entry with a 6-hour duration
• Ensure the cache key includes repoName to avoid cross-repository collisions
• Preserve cache bypass behavior for manual refresh actions

Please assign this issue to me.

Thank you!

[github-actions]

Assigned to @Sammmyyyyyyy! You have 7 days to open a PR. No PR after 7 days = auto-unassigned so others can pick it up. Good luck!