Laravel Scout + Meilisearch v1.11+ Production Docker Compose & Horizon Queue Tuning

Preventing Synchronous Eloquent Saves From Flooding Meilisearch Tasks (MeiliSizer Guide)

Out of the box, saving 5,000 Eloquent models in a loop dispatches 5,000 tiny 1-document indexing tasks in Meilisearch, creating a massive task queue backlog. Enabling SCOUT_QUEUE=true on a dedicated Horizon supervisor and wrapping bulk updates in Model::withoutSyncingToSearch() + searchable() chunks collapses 5,000 tasks into 5 ultra-fast LMDB commits.

Production docker-compose.yml & config/scout.php Index Settings

# docker-compose.yml — Isolated Meilisearch v1.11 Service on 16GB KVM
services:
  meilisearch:
    image: getmeili/meilisearch:v1.11
    restart: unless-stopped
    environment:
      MEILI_ENV: production
      MEILI_MASTER_KEY: ${MEILISEARCH_KEY}
      MEILI_MAX_INDEXING_MEMORY: 4Gb
      MEILI_NO_ANALYTICS: "true"
    volumes:
      - meili_lmdb:/meili_data
    ports:
      - "127.0.0.1:7700:7700"
    deploy:
      resources:
        limits:
          memory: 10G

MeiliSizer RAM & NVMe Tier Matrix for LARAVEL-SCOUT-DOCKER

MeiliSizer Production Checklist: LARAVEL-SCOUT-DOCKER

• Run php artisan scout:sync-index-settings inside your deployment pipeline before running migrations or bulk imports.
• Set SCOUT_QUEUE=true and assign a dedicated Redis Horizon queue ("search-sync") with chunk sizes of 500–1,000 records.
• Define custom toSearchableArray() on every Eloquent model so sensitive columns and giant JSON blobs are never serialized.
• Bind port 7700 to 127.0.0.1 or a private VPC interface; expose public search only via Caddy/Nginx TLS with scoped keys.

Load Workload in MeiliSizer Matrix →