inotify-max-user-instances-exhausted-by-docker-shims
Root cause: each Docker container shim consumes ~1 inotify instance; 100+ containers exhaust the default kernel limit of 128, causing brainstorm server (and other Node.js watchers) to fail with EMFILE. Fix: create /etc/sysctl.d/99-inotify.conf with fs.inotify.max_user_instances=512 and apply with sysctl —system. Symptom is misleading — the error appears as a file-descriptor issue, not a container count issue.
Related
- inotify-limit-saturation-from-docker-containers
- docker-shims-exhaust-root-inotify-instances
- inotify-instances-exhausted-by-docker-shims
- inotify-limit-saturated-by-docker-container-shims
- emfile-too-many-open-files-on-fswatch-nodejs-as-root-on-vps
- ast-grep-npm-install-overwrites-system-sg-newgrp-symlink
- systemd-restart-on-failure-misses-clean-idle-timeout-exits
- email-alert-from-address-defaults-to-workflow-user
- pine-script-default-rr-ratio-is-negative
- hypervisor-steal-time-misread-as-process-cpu
- docker-build-cache-accumulates-indefinitely-without-cron
- dockerfile-useradd-chown-before-copy-doubles-layer-size
- prometheus-cpu-steal-alert-thresholds-for-hostinger-vps
- docker-daemon-json-log-rotation-already-configured-system-wi
- hostinger-applies-kernel-upgrades-via-silent-vps-reboot
- mass-container-reinit-post-reboot-mimics-attack-pattern
- r-dash-pilot-launch-criterion-10x5x15x30x0-wafh-window
- rdash-env-var-naming-drift-breaks-prod-boot
- dependabot-enable-triggers-immediate-pr-storm