176 New Projects. An Unplanned Stress Test.

176 new projects on VibeNest in roughly 12 hours. I'd have preferred to celebrate with a beer rather than disk cleanup.

Alongside the surge came platform abuse. Unwanted workloads consumed capacity, one application host ran out of space, and deployments failed.

That failure is on us. We're sorry for the disruption. While we address it, deployments and application availability may still be affected.

Cloudflare provides DDoS protection. But traffic filtering alone cannot solve our deployment and storage problem: projects keep consuming resources after their launch requests have gone through. We need controls inside VibeNest too.

We reclaimed capacity while preserving project databases and volumes. Checks afterward found no regressions caused by the cleanup. Then new builds filled the host again. The problem needed more than another cleanup.

Here's what we're working on:

• Project behavior: strengthen analysis of how projects interact with external services and consume resources, so we can identify and block abusive workloads.
• Capacity checks: account for disk space alongside memory and CPU before placing new builds.
• Image retention: define which images to keep for recovery and when to remove them. Stopping a project doesn't automatically remove its images.

The abusers also gave us an unplanned stress test. No need to recruit testers; they found us.

That pressure exposed real limits in build placement, storage, and abuse prevention. It's useful to find them at this stage. Now we have concrete failure cases to fix and verify before we grow further.

Thank you for bearing with us. Keeping your projects running comes first. The beer can wait.