Skip to main content
Sponsor @cloudposse

Decide on RDS Aurora DB Cluster Requirements

Problem​

Requirements for Amazon Aurora DB clusters deployed to each active compute environment need to be outlined before an Amazon Aurora component is configured and deployed

Context​

Amazon RDS provides MySQL and PostgreSQL-compatible relational databases that are built for the cloud with greater performance and availability at 1/10th the cost of traditional enterprise databases with the simplicity and cost-effectiveness of open source databases. RDS Aurora features a distributed, fault-tolerant, self-healing storage system that auto-scales up to 128TB per database instance. It delivers high performance and availability with up to 15 low-latency read replicas, point-in-time recovery, continuous backup to Amazon S3, and replication across three Availability Zones.

Amazon Aurora DB clusters (See: Decide on RDS Technology and Architecture)

Known Limitations​

Considered Options​

Create a standardized Aurora DB cluster based on the current use case:

Current Infrastructure Requirements​

RDS Aurora Replication​

RDS aurora replication happens at the filesystem layer versus the conventional database layer. It’s actually a shared filesystem. Hitting the read replicas hard can still impact the masters since they are using the shared filesystem.

Because the cluster volume is shared among all DB instances in your DB cluster, minimal additional work is required to replicate a copy of the data for each Aurora Replica. https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/Aurora.Replication.html

RDS Serverless v1 vs v2​

Using serverless could be more costly that using regular RDS aurora due to not having enough options for CPU.

Serverless v1 offers more granular scaling units. Only operating in a single availability zone. Serverless v1 only supports up to v10 of Postgres (v10 will be sunset by Postgres on November 10, 2022).

Serverless v2 offers multi-AZ, so that DB subnets can be across multiple availability zones. Supports Postgres 12+.

https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/aurora-serverless-v2.upgrade.html

Future Aurora DB Cluster Requirements​

Because the RDS Service documentation on Aurora DB clusters recommends deploying at least one additional Aurora DB cluster instance to an Availability Zone other than where the primary instance is located in order to ensure High Availability, and since 3 AZs are used by the primary AWS region, it is recommended that 3 instances are deployed per Aurora DB cluster when High Availability is needed.

  • Explain how many instances exist in the cluster (or per region, if this is a global cluster)

  • Explain whether the cluster is global or regional, and reference Decide on RDS Technology and Architecture

  • Explain how many secondary DB clusters should exist, if this is a global cluster

Lastly, database storage encryption, deletion protection and cloudwatch logs exports should be enabled as a best practice.

This, in addition to any of the requirements outlined in v1 Infrastructure Requirements, should be captured in the following table.

Setting​

Aurora DB cluster Engine
Aurora DB cluster Instance Family
Number of Aurora DB cluster Instances: 1 for all environments except for prod, 3 for prod (or 2 when < 3 AZs are available)
Regional or Global DB Cluster
Security-related settings
Storage Encryption enabled
yes

Other Considerations​

Consequences​

Create an Aurora DB Cluster component and tune it to the outlined requirements.

References​