WAL SQLiteEdge SQLite Archive
BP-04 // RAILS 8 NO-PAAS

Rails 8 + SolidQueue, SolidCache & SolidCable on Kamal 2 NVMe VPS

SQLiteEdge Rails 8 blueprint: isolating primary, queue, cache, and cable .sqlite3 WAL files on one Kamal 2 NVMe host to eliminate managed Redis.

MODELED DESIGN TARGET
4x Isolated WALs
READ PATH / RPO TARGET
0.42 ms
REFERENCE NVME PROFILE
$10.49–$48 NVMe

Rails 8 made SQLite a first-class production citizen by shipping SolidQueue, SolidCache, and SolidCable alongside automatic WAL/IMMEDIATE patches in ActiveRecord. The golden rule on a Kamal 2 single-server deployment is four separate SQLite files (`production.sqlite3`, `queue.sqlite3`, `cache.sqlite3`, `cable.sqlite3`) so high-churn cache evictions and job polling never contend with customer checkout writes.

In our SQLiteEdge BP-04 reference guide, every configuration rule below follows official SQLite 3.46 documentation and public vendor specifications retrieved 2026-11-12.

SQLiteEdge Architecture Rules for Rails 8 SolidQueue, SolidCache & Kamal Multi-SQLite

  1. Declare four separate database entries under production: in config/database.yml—each SQLite file gets its own independent single-writer lock and -wal journal on the NVMe bus!
  2. Bind-mount /var/lib/kamal-storage:/rails/storage on the host NVMe drive so container deploys never wipe persistent SQLite volumes.
  3. Size SolidCache max_size generously on fast local NVMe (e.g., 20 GB – 60 GB disk cache) instead of paying $120/mo for tiny 2 GB managed Redis instances.
  4. Back up ONLY production.sqlite3 (and optionally queue.sqlite3) via Litestream—never waste S3 bandwidth replicating ephemeral SolidCache entries.

SQLiteEdge Reference Snippet: config/database.yml (Rails 8 Multi-SQLite Production Layout)

config/database.yml (Rails 8 Multi-SQLite Production Layout) 
default: &default
  adapter: sqlite3
  pool: <%= ENV.fetch("RAILS_MAX_THREADS") { 8 } %>
  timeout: 5000
  pragmas:
    journal_mode: wal
    synchronous: normal
    mmap_size: 4294967296
    cache_size: -524288
    journal_size_limit: 67108864

production:
  primary:
    <<: *default
    database: storage/production.sqlite3
  queue:
    <<: *default
    database: storage/production_queue.sqlite3
  cache:
    <<: *default
    database: storage/production_cache.sqlite3
  cable:
    <<: *default
    database: storage/production_cable.sqlite3

SQLiteEdge Modeled Profile & Tariff Table (BP-04)

Public vendor prices retrieved 2026-11-12 (promotional and renewal rates may change at checkout); throughput figures reflect theoretical single-queue sizing models rather than unarchived lab claims.

Infrastructure ComponentHeroku / AWS Managed StackRails 8 + Kamal on Local NVMe
Web + Worker Compute (16GB)$250/mo (Dynos / ECS)Included in $10.49–$48/mo NVMe VPS
Primary Database (50 GB)$165/mo (Managed Postgres)production.sqlite3 ($0 extra)
Job Queue + Cache Store$115/mo (Managed Redis)SolidQueue + SolidCache on NVMe ($0)
Total Monthly Infrastructure$530.00/mo$16.49 – $54.00/mo (incl. $6/mo DR)

SQLiteEdge Spec-Matched NVMe & S3 Targets for BP-04

Open this BP-04 workload preset inside the SQLiteEdge PRAGMA Studio →

SQLiteEdge Engineering FAQ for BP-04

  • Will SolidQueue polling starve my primary user transactions in Rails 8?

    Not when queue.sqlite3 is configured as a separate database file in database.yml! SQLite locks are per-file, so SolidQueue workers can hammer production_queue.sqlite3 while checkout writes enjoy 100% uncontended access to production.sqlite3.

  • Should I choose Hostinger EPYC KVM or Vultr High Frequency for Kamal 2?

    Start on Hostinger KVM 4 (4 vCPU / 16GB / 200GB NVMe at $10.49/mo) for most SaaS apps; upgrade to Vultr 3.5GHz+ High Frequency when SolidQueue processes >1,500 jobs/second or Ruby single-thread response time becomes critical.