AWS RDS - keshavbaweja-git/guides GitHub Wiki
Introduction
- Managed Relational Database Service
- Six database engines - Amazon Aurora, MySQL, MariaDB, Oracle, Microsoft SQL Server, and PostgreSQL.
- RDS handles routines database tasks - provisioning, patching, backups, restore/recovery, failure detection and repair.
- Mutli-AZ deployment offers
- High availability, reliability
- Automated failover from primary database to synchronously replicated secondary database.
- Read replicas - scale out beyond the capacity of a single database instance for read heavy database workloads
Database Instance Type
- Standard
- Memory Optimized
- Burstable Performance
Performance
General Purpose SSD Storage
- Consistent baseline of 3 IOPS per provisioned GB
- Baseline performance of 3000 IOPS for 1 TiB volume
- Minimum baseline performance of 100 IOPS
- Maximum baseline performance of 16,000 IOPS for 5.34 TiB and above
- Supports IO Bursts up to 3000 IOPS
- Volumes below 1 TiB in size also have ability to burst to 3,000 IOPS for extended periods of time
- Suitable for a wide range of database workloads
- Can experience larger latency in scenarios of read replica creation, multi-az conversion, and db snapshot restoration. For minimum latency in these scenarios, provisioned IOPS instance is recommended
Provisioned IOPS SSD Storage
https://docs.aws.amazon.com//AmazonRDS/latest/UserGuide/CHAP_Storage.html#USER_PIOPS
- Fast, predictable and consistent IO performance
- Suitable for IO intensive OLTP database workloads
Scalability
- Push Button Compute Scaling - Scale up or down in minutes, up to a maximum of 96vCPUs and 256 GB RAM
- Easy Storage Scaling
- Amazon Aurora will automatically scale up to a maximum of 64 TB, maximum size can be user configured
- Other database engines can scale up to 16 TB on- the-fly with zero downtime.
Availability and Durability
Automated Backups
- Turned on by default
- Backs up database and transaction logs, stored for a user specified retention period
- Restore database instance to any second during your retention period, up to last five minutes
Database Snapshots
- User initiated database backups,
- Stored in S3, retained until explicitly deleted
Security
- Encryption at Rest and in Transit
- Encrypt data using keys managed by Amazon KMS.
- With Amazon RDS encryption enabled, data in underlying storage, automated backups, snapshots and read replicas are all encrypted
- Transparent Data Encryption (TDE) is supported for Oracle and Sql Server.
- TDE for Oracle is integrated with CloudHSM (single tenant Hardware Security Module)
- SSL is used to encrypt data in transit.
Monitoring and Metrics
- Amazon RDS provide CloudWatch metrics for you database instance at no additional charge.
- Key operational metrics like compute, memory, storage utilization, IO activity and instance connections are available.
- Enhanced monitoring provides access to over 50 metrics
Storage metrcis
-
IOPS
-
Throughput
-
Latency
-
Queue Depth: avg queue depth reported in 1 min intervals. Time spent in queue is a component of latency and is not reported as a metric
-
Event Notifications
- Database events can be notified via email or SMS through Amazon SNS (Simple Notification Service)