Nuxt SSR webfrontend memory leak causes OOM crash on idle all-in-one Docker deployment

Are you using our SaaS platform (Baserow.io) or self-hosting Baserow?

Self-hosted

If you are self-hosting, what version of Baserow are you running?

Baserow version: 2.0.6 (all-in-one Docker image baserow/baserow:2.0.6)

If you are self-hosting, which installation method do you use to run Baserow?

Deployment: Dokku (Docker-based PaaS) on a VPS with 11 GB RAM

What are the exact steps to reproduce this issue?

The Nuxt.js webfrontend process in the all-in-one Docker image has a memory leak that causes the container to be OOM-killed even with zero user traffic. The Node.js process grows at approximately 10-15 MB/minute while completely idle, eventually exhausting the container’s memory limit.

Environment

  • Baserow version: 2.0.6 (all-in-one Docker image baserow/baserow:2.0.6)
  • Deployment: Dokku (Docker-based PaaS) on a VPS with 11 GB RAM
  • Container memory limit: 3 GB (–memory=3g)
  • Host OS: Debian (Linux)
  • External Postgres and Redis (Dokku-managed services, not embedded)

Relevant env vars:
BASEROW_AMOUNT_OF_GUNICORN_WORKERS=1
BASEROW_AMOUNT_OF_WORKERS=1
BASEROW_RUN_MINIMAL=yes
BASEROW_CELERY_BEAT_STARTUP_DELAY=60
SYNC_TEMPLATES_ON_STARTUP=false
BASEROW_TRIGGER_SYNC_TEMPLATES_AFTER_MIGRATION=false

Hey @biobaser, sorry for the delayed response! There’s a chance this issue will be resolved with the 2.1 release today or tomorrow, so could you please try again once it’s out?