Runtime and processes
Check a customer-managed database server where memory, durable storage, backups, network exposure, authentication, maintenance windows, and restore testing drive the design.
Database server VPS
Run your database on a server you administer. Size memory and storage for your data, then plan access controls, backups and restore testing.
Hosting options
Self-managed virtual servers with administrator access, NVMe disks, shared vCPU, dedicated IPv4 and IPv6. RAM uses ballooning. Backups are sold separately.
See VPS hosting →02Canadian SFTP storage for your backup software. Check physical separation from your source server before treating a copy as off-site.
See Backup storage →03Register or transfer a domain and manage DNS. Registration and renewal prices depend on the extension.
See Domains →04Tell us what you need to host and when it needs to be ready.
See Ask about your project →Your requirements
Check a customer-managed database server where memory, durable storage, backups, network exposure, authentication, maintenance windows, and restore testing drive the design.
Identify databases, uploads, volumes, configuration and secrets that cannot be rebuilt from the application code. Back them up and test a restore.
List the ports that must be public and those that should remain private. Configure authentication, TLS and firewall rules before accepting traffic.
Check how the service starts after a reboot, where logs are stored and who applies updates. Test a deployment and rollback without changing live customer data.
Field guide
A database server spends its budget on memory for caches and working sets, plus storage latency for every write that must reach disk. For InnoDB or PostgreSQL, the buffer pool or shared buffers setting is the main memory dial, and it should leave room for connections, sort and join work, and the operating system cache.
Connections cost memory each, so an application pool multiplied across several app servers can exceed the database limit or exhaust RAM through per connection buffers. Keep the database port off the public internet, require authentication for every account, and test a full restore on a schedule, since an untested backup is a hope, not a recovery plan.
Set the main cache to hold your active working set, then confirm with cache hit ratios in the database status output. A working set that outgrows the cache shows up as constant disk reads.
Add up connection pools from every application server and cron job before raising the database connection limit. Each connection carries its own memory overhead.
Restore the latest backup to a separate instance and run the application's health checks against it. Record how long the restore took; that number is your real recovery time.
Planning
Confirm the current product specifications and service responsibilities. Record who can approve a change and how to recover the service if it goes wrong.
Use the detailed checklist to plan setup, access, migration and recovery with your team.
Download PDF →Send us the software you use, your storage needs and your launch or moving date.
Ask about hostingYour current site stays online until your new one is ready. You approve the copy first, and if anything is not right we switch you back right away. The price you pick is the price you renew at.