Managed database or self-hosted: which should a small D2C brand choose?

Comparing managed databases like Amazon RDS with running your own on a server: backups, patching, failover, cost and which suits a small D2C team.

Anadoris Digital · · 4 min read

Archive storage
Photo: Samuel Zeller / Unsplash (CC0)

Your database holds the things you cannot lose: orders, customers, inventory, invoices. Whether you run it yourself or let a provider manage it is one of the most consequential infrastructure decisions a small brand makes, even if it rarely gets much thought.

Self-hosting means installing PostgreSQL or MySQL on a virtual server you manage. Managed means a service such as Amazon RDS, Aurora or a database offered by your hosting platform, where the provider runs the engine and you use it.

What a managed database handles

  • Automated daily backups and point-in-time recovery within a retention window you choose.
  • Minor version patching during maintenance windows you set.
  • Optional standby copies in another availability zone for automatic failover.
  • Monitoring of storage, memory and connections in one console.
  • Encryption at rest with a few settings.

What self-hosting asks of you

  • Writing, scheduling and checking backups, and copying them off the server.
  • Testing restores so you know backups actually work.
  • Applying security updates to the database and operating system.
  • Watching disk space; a full disk can stop the database entirely.
  • Planning failover if the server dies.

None of this is exotic, but all of it must happen every month, forever. In small teams it tends to slip until the day it is needed, which is usually the worst possible day: a corrupted table during a sale, or a failed disk just after the developer who set things up has left.

For a small brand, the honest question is not whether your team could do these tasks, but whether they will still be doing them on a busy Friday six months from now, when the person who set it up has moved on. Managed services exist mainly to remove that dependency on memory and goodwill.

Cost: compare the whole picture

A managed database usually costs more than a bare server of the same size. That gap pays for the work above. When comparing, include the hours someone spends on operations and the cost of a lost day of orders. Check current AWS pricing for the instance size, storage and backup retention you need, and consider smaller instance sizes for staging.

Settings that matter on a managed database

  • Backup retention long enough to notice a problem before it rolls off.
  • Deletion protection turned on for production.
  • Database not publicly reachable; only your apps can connect.
  • Encryption at rest enabled when the database is created.
  • Alarms for low free storage, high CPU and too many connections.
  • A named owner who checks the maintenance window and upgrade notices.

When self-hosting can make sense

Self-hosting is reasonable for throwaway experiments, local development, or teams with genuine database skills and time. It can also suit a tiny internal tool where losing some data is acceptable. For order and customer data, the managed route is almost always the safer default for a small brand. If you do self-host, treat backups and restore tests as non-negotiable, store backups in a separate account or region, and write down the exact steps to rebuild the server from scratch.

Performance and connections during sales

Databases are often the first part of a D2C setup to struggle on a sale day. More shoppers mean more webhooks, more stock checks and more reports running at once. The common failure is not raw speed but connections: each app instance or serverless function opens its own connection, and the database reaches its limit.

On a managed service you can add a connection pooler such as Amazon RDS Proxy, scale up the instance size before a known event, and add a read replica so reports and dashboards do not compete with order processing. Self-hosting can do the same, but you configure each piece yourself. Whichever route you take, look at slow query logs a week before a sale and add missing indexes while things are calm. Also check that heavy jobs such as exports, report generation and bulk updates are scheduled away from the hours when your campaign goes live, so they do not compete with shoppers for the same database.

The same care applies to who and what can reach your data:

  • Separate database users for each app, each with only the rights it needs.
  • Passwords held in a secrets manager and rotated when people leave.
  • Read-only access for analysts and dashboards.
  • No copies of production data on laptops or in shared drives without a clear reason.

Where to start

Find out where your databases run today and who backs them up. If any production database is self-hosted with no tested restore, either set up and test proper backups this month or plan a move to a managed service. If you already use RDS, confirm deletion protection, backup retention and private access are set correctly. Then schedule a restore test into a separate instance, time how long it takes, and write the steps down. That one exercise tells you more about your real risk than any checklist.

Related service: AWS & Cloud Management · More on this topic: AWS & Cloud Management articles

Common questions

What is the difference between RDS and Aurora?

Both are AWS managed relational databases. RDS runs standard engines such as PostgreSQL and MySQL on managed instances. Aurora is AWS's own engine compatible with PostgreSQL and MySQL, with a different storage design and features. Compare current pricing and features for your workload before choosing.

Can I move from a self-hosted database to RDS?

Yes. Small databases can be moved with a dump and restore during a short maintenance window. Larger or busy ones can use AWS Database Migration Service to copy data while the old one keeps running. Test the move on a copy first.

Do managed databases ever lose data?

Managed services greatly reduce the risk but do not remove it. Someone can still delete a table by mistake, or a bug can corrupt records. Point-in-time recovery, deletion protection and occasional restore tests are what protect you.

Should I use a NoSQL database like DynamoDB?

It suits specific patterns such as high-volume lookups by key, sessions or event logs. For orders, customers and inventory that need relationships and reporting, a relational database is usually the easier choice for a small team.

Your next chapter

Let’s build what comes next.

A new website. A better system. A simpler way to work. Tell us where you want to begin.

Discuss your project