Amazon RDS consulting for always-on databases

Amazon RDS (Relational Database Service) runs a SQL database for you on AWS. It supports Aurora, PostgreSQL, MySQL, MariaDB, SQL Server, Oracle and Db2. AWS manages the hardware, patching, automated backups and failover, while you manage the schema, queries and access. Most of our clients use it behind web apps, customer portals and internal tools.

Aurora is part of RDS but uses its own storage layer that keeps six copies of your data across three Availability Zones and grows on its own. It suits busy or spiky apps, many read replicas, or a database that must span Regions with Aurora Global Database. Standard RDS for MySQL or PostgreSQL is often cheaper for small, steady databases. We compare both on your real load before deciding.

On EC2 you get full control of the server, and you also own every patch, backup and failover script. RDS takes those chores off your plate in exchange for less low-level access. We suggest EC2 only when you need an extension, plugin or OS agent that RDS doesn’t allow; otherwise RDS is less work to keep healthy.

RDS is relational: tables, joins, transactions and SQL. DynamoDB is a key-value and document database that scales to very high traffic when you know your access patterns up front. Apps with reports, joins and changing questions usually belong on RDS; carts, sessions and event streams at large scale often fit DynamoDB better.

Automated backups can be kept from 1 to 35 days, and within that period you can restore to any second up to roughly the last five minutes. Manual snapshots stay until you delete them. Backup storage up to the size of your total database storage in a Region costs nothing extra; beyond that RDS for MySQL and PostgreSQL charge $0.095 per GB-month in US East.

Multi-AZ keeps a synchronous standby in another Availability Zone, and RDS fails over to it automatically while your endpoint name stays the same. With the classic one-standby setup you can’t send queries to the standby. A read replica is an asynchronous copy you can query for reports; RDS for MySQL, MariaDB and PostgreSQL allow up to 15. Many production setups use both.

We copy the full database while the old server keeps running, then keep it in sync with change data capture through AWS DMS or the engine’s own replication. When the copy is caught up and tested, we stop writes, let the last changes land, and point the app at RDS in a short agreed window, with the old server kept ready in case we need to roll back.

It depends on the engine, instance class, storage and whether you use Multi-AZ. In US East, RDS for PostgreSQL on db.t4g.medium is $0.065 per hour Single-AZ or $0.129 per hour Multi-AZ, and gp3 storage is $0.115 or $0.23 per GB-month. A small Multi-AZ database with 100 GB comes to about $117 a month before data transfer. Reserved Instances lower the hourly rate for steady use.

Yes. Zoho Analytics can import tables from RDS on a schedule using a read-only user. For live business events we usually put Lambda and an API in between, for example a Zoho CRM webhook creating an account row, or new orders in RDS becoming Zoho Books invoices. We avoid opening the database itself to the internet.

Yes, for up to seven days at a time; after that AWS starts it again automatically so it can apply required maintenance. You stop paying instance hours while it is stopped, but storage and backups are still billed. It works well for test and staging databases, and we often schedule it.