Amazon Aurora¶
TLDR¶
AWS Created custom DB tech, very high performan. Useful for global and serverless relational database applications.
Features¶
- not open source
- Failover is instantaneous HA Native
- MySQL and Postgres
- global
- self-healing
- Fault Tolerant
- storage automatically grows in increments of 10Gb up to 128 Tb
- autoscaling up to 64 TB per instance
- is a cluster
- multi AZ
- each AZ has a copy of the cluster.
- 5x faster than MySQL.
- 3x fast than Postgres.
- up to 15 replicas.
- faster replication.
- 20% more cost than RDS
- automated patching with zero downtime
- backtrack restore to any point in time without backups (up to 72 hours)
High Availability & Read Scaling¶
- 6 copies of your data across 3 AZ
- 4 copies out of 6 needed for writes
- 3 copies out of 6 needed for reads
- self-healing with peer-to-peer replication.
- Storage is striped across 100s of volumes.
- One aurora instance takes writes (master)
- automated failover for master in less than 30 seconds.
- Master + up to 15 read replicas to serve reads
- support for cross region replication.
- Cross region read replicas:
- useful for Disaster Recovery
- simple to use
Aurora Db Cluster¶
- Reads replicas are behind auto-scaling
Primary DB¶
- read and write
- only one per cluster
- if fails a replica will be promoted
Aurora Replica¶
- same storage as primary
- multiple per AZ possible
- only read
- reader endpoint does load balancing
- can enable replica autoscaling
Custom Endpoints¶
- subset of aurora instances as custom endpoint
- once a custom endpoint is enabled, reader endpoints no longer works.
- (e.g., larger instance type for analytical queries)
Aurora Serverless¶
- automated db instantiation and autoscaling
- useful for infrequent, unpredictable and intermittent workloads
- no capacity planning needed.
- Pay per second
- client talks to aurora proxy fleet.
Aurora Multi-master¶
- use for immediate failover (High Availability)
- all nodes are R/W vs. promoting a RR (Read Replica) as a new master,
- there is no master
- replication between all nodes
Global Aurora Database¶
- 1 primary region for R/W
- up to 5 secondary read regions
- up to 16 read replicas per secondary read region.
- Less than 1 second for replication across region
- helps for decreasing latency worldwide.
- Promoting another region for (DR) has an RTO of < 1 minute.
Aurora Machine Learning¶
- ML based predications to your apps via SQL
Use cases¶
- Fraud Detection
- add targeting
- Sentiment analysis (AWS comprehend)
- product recommendations
Services¶
Backups¶
Automated¶
- 1 to 35 days (cannot be disabled)
- point in time recovery in that timeframe
Manual¶
- triggered by user
- keep as long as user wants
Restore¶
- into new database
Cloning¶
- create a new aurora cluster from an existing one
- faster than snapshot & restore
- cost-effective
- useful to create staging