Context

This project (favorite) was deployed to Vercel at fav.jpdiaz.dev. After going live, it triggered a massive spike in resource usage, which resulted in Vercel pausing my entire account due to quota overages on the free plan.

Symptoms

• All Vercel projects were paused with message:

  “This deployment is temporarily paused. Upgrade to resume service.”

• Error on domain: 404: DEPLOYMENT_NOT_FOUND
• All limits were massively exceeded (based on Usage PDF from Vercel dashboard):

Root Cause

From Vercel’s Observability dashboard, the overwhelming usage came from this route:

• Route: /_render
• Requests: 112,000+ in <12 hours
• Outbound transfer: 2.38 GB
• Inbound transfer: 91.28 MB

This suggests that /_render is either:

• An internal server-side rendering (SSR) endpoint exposed publicly
• Not cached
• Not rate-limited
• Possibly being hit by bots, crawlers, or unintentional traffic loops

What happened next

• I deleted the project from Vercel (June 16)
• Usage stopped growing immediately
• Confirmed via dig that DNS now points to Netlify (not Vercel)
• Ported portfolio to: https://jpdiaz.netlify.app

Next steps (to be handled in future issues)

• Analyze what triggers /_render and why it was exposed
• Add SSR caching or convert to static generation where possible
• Block bots and crawlers from hitting dynamic endpoints
• Add logging to monitor endpoint traffic
• Add rate limiting or edge protection for production

This issue is only for documenting the root cause. Future issues will address mitigation.