Environment

• Operating System: Linux (Google Cloud Run)
• Node Version: v24
• nuxt/cli version: 3.32.0
• Package manager: npm
• Nuxt Version: 4.2.2
• Nitro Version: 2.12.9
• Builder: vite 7.3.0
• Config: $development, $env, $production, app, compatibilityDate, components, content, css, devServer, devtools, eslint, experimental, image, modules, nitro, pages, routeRules, runtimeConfig, scripts, typescript, vite
• Modules: @vueuse/nuxt 14.1.0, @nuxt/image 2.0.0, @nuxtjs/tailwindcss 6.14.0, @nuxt/test-utils/module 3.20.1, @nuxt/content 3.8.0, @nuxt/eslint 1.10.0, @nuxt/scripts 0.13.1, @nuxt/icon 2.1.0

Reproduction

Cannot provide minimal reproduction - occurs only in production SSR under real traffic.

Describe the bug

Continuous memory leak in Nuxt 4.2.2 SSR application. Memory grows until OOM crash.

Additional context

Heap Profile Comparison (Before vs After ~6 hours of traffic)

Total heap: ~111 MiB → ~843 MiB (7.6x growth)

Key Observations

1. Vue Router shows highest growth - tokensToParser increased 18x
2. Data fetching accumulates - useAsyncData/useFetch only appear after growth
3. All functions grow proportionally - indicates entire SSR context is not being released

• Aware of Memory leak in Vue #33644, updated to 4.2.2, issue persists
• Memory grows from ~400MB to 2GB+ over several hours

What we've tried

1. Audited all application code for module-scope state leaks
2. Verified no module-scope useNuxtApp(), useRoute() calls
3. No improvement from application-side fixes

Logs