In particular, I have a question, what is the recommended way to organize a failover in AWS. It will also be useful to understand typical AWS failures in order to organize Application HA (high availability). Thus, application architecture (using AWS): This is a more / less typical web application architecture in AWS
- There is a route 53 that permits the ip of some ELB.
- There is a public subnet in which there is an ELB, and it directs traffic to web servers to a private VPC;
- In a private subnet route, go: Web servers → ELB-> Application servers;
- Application servers write data to Multi-AZ RDS.
The main drawback of such a deployment is that services are active in one AZ, because in a Multi-AZ deployment, Amazon RDS automatically backs up a synchronous backup to a different availability zone. So, the master is in only one AZ, and services in another AZ are not allowed for writing to RDS, because it is in standby mode.
Two questions:
- What is the best way to implement HA for such a deployment?
- What are the common AWS errors (if one AZ is unavailable, does this often happen only with some services (for example, other problems VPC / EC2 / EBS?) Or, as a rule, all AZ services are not available)?
HA considerations for this approach:
- RDS. AWS: " Amazon RDS , Multi-AZ. , ...". , AWS RDS.
- / AZ. Route53 Active AWS AZ. RDS ( , RDS AZ, AZ)?
AZ , IP- (, IP- ). 1 - EC2 AZ, ( AZ (s)).