What is a cloud database?
A cloud database is a database that runs in and is managed through the cloud — often delivered "as a service" (DBaaS) so the provider handles the servers, patching, backups and scaling while you focus on your data. It stores and organises information exactly like a database always has, but the operational work of keeping it running — provisioning hardware, applying updates, taking backups, planning for failure — shifts partly or entirely to the provider. That shift is why cloud databases have become the default choice for new applications, and it is a concept every cloud certification touches early. Here is what a cloud database actually is, how managed compares with self-managed, and where the major providers’ offerings fit.
The plain idea: hosted and managed, not just installed elsewhere
You could always run a database "in the cloud" in the loosest sense — install database software on a cloud server and administer it yourself, much as you would on a physical machine. A cloud database usually means something more specific: a managed service where the provider runs the database engine for you, and you interact with your data through a connection endpoint rather than logging into the underlying machine at all.
The provider takes on the operational load that used to sit with a database administrator — provisioning capacity, applying security patches, running backups, monitoring health and handling hardware failure — in exchange for a usage-based fee. You are still fully responsible for your schema, your queries and your data itself; the provider’s responsibility is keeping the engine running underneath it.
Managed vs self-managed
This split mirrors the shared-responsibility idea that runs through cloud computing generally: the provider secures and maintains the underlying infrastructure, and you remain responsible for what you put on top of it — your data, your access controls, your query design. A fully managed cloud database pushes more of that line toward the provider than installing database software yourself on a rented server would.
Self-managed still has a place — some teams need control over a specific engine version, an unusual configuration, or fine-grained tuning that a managed service does not expose. But for most applications, the operational burden a managed database removes — patching at 2am, restoring from backup after a failure, planning capacity by hand — is exactly the kind of undifferentiated work cloud computing exists to take off your plate.
Relational vs NoSQL cloud databases
Cloud databases come in the same two broad families as databases generally. Relational (SQL) databases organise data into structured tables with defined relationships, and suit data with a consistent, well-understood shape — customer records, orders, financial transactions. NoSQL databases trade that rigid structure for flexibility and horizontal scale, suiting data that is less uniform or needs to scale across many servers — session data, catalogues, high-volume event data.
Our dedicated guide on relational versus NoSQL databases covers the distinction in full; the point for this article is that cloud providers offer managed versions of both, so choosing "a cloud database" always means choosing which family fits your data first, and which specific managed service second.
The benefits, briefly
The case for a managed cloud database over running one yourself comes down to a few consistent advantages: managed operations (patching, backups and monitoring handled for you), scalability (capacity adjusted as demand grows, without a manual migration), and high availability (built-in replication and failover options that would take real engineering effort to replicate by hand). None of these are unique to any one provider — they are the general case for choosing "managed" over "self-managed" in the cloud.
Cloud databases across the major providers
Every major provider offers managed relational and NoSQL databases under its own names. On AWS, Amazon RDS covers managed relational engines and Amazon DynamoDB is the managed NoSQL option — our dedicated explainer on what Amazon RDS is covers the relational side in depth. Azure’s equivalents are Azure SQL Database (managed relational) and Azure Cosmos DB (managed, globally distributed NoSQL). Google Cloud offers Google Cloud SQL for managed relational databases, alongside its own NoSQL options. As with cloud servers generally, the underlying concept is the same across all three — the differences are mostly in naming and the surrounding ecosystem.
Where this fits in certification study
Databases are a recurring topic across cloud certifications because so many applications depend on one. Associate-level exams such as AWS Solutions Architect Associate and Azure Administrator Associate expect you to know when to choose relational versus NoSQL and how managed services reduce operational overhead, while a data-focused credential like AWS Data Engineer Associate goes considerably deeper into designing and operating data workloads on managed services. Understanding the managed-versus-self-managed distinction here is a solid base for any of those paths.
Original practice questions, timed mock exams and revision notes. No card, nothing to pay.