Aurora

  • Proprietary database by AWS
  • Data is replicated to 6 Replicas in 3 AZ
  • Auto healing capability
  • Multi AZ
  • Auto scaling read-replica
  • Storage from 10GB to 64TB
  • Aurora Global database for DR and Latency Uses
  • Need to provision EC2 Instance and EBS Volume
  • Using Aurora Serverless does not require instance and volume
  • 5x faster than regular RDMS
  • Support MAX 15 Read Replicas
  • We can create custom endpoint for
  • Production Database
  • Reporting queries
  • Built in reader-endpoint can be used for distribute the traffic between read-replicas
  • If DB Primary Instance fails, it create a new DB Instance in the same AZ as the Original Instance and done by Best Effort basis
  • Read Replicas can be used to avoid un necessary downtime
  • With binlog replication, another aurora cluster can be used as read replica. Aurora cluster can be same or different region
  • Aurora auto scaling policy can be used to adjust performance
  • Based of number of connections (when the lambda function is using the database)