AWS CostOptimization intiatives

1. SQL Server backup:

Problem: Today, the SQL backups are being storged in backup volume or in similar place in same EC2 server. First this doesnt help for recovery and Second this leads to duplication of storage as AWS Snapshots take backup of backup.

Solution: Use storage gateway to offload the backups to reduce the local EBS storage being used by backups. The gateway can be implemented in 2 ways, 1-In a OU (group of aws accounts) to share between various account to save gateway device cost or 2-In each aws account if there is a business need or it there is too-much data being traversed over VPC's

1. AWS Opensearch GP2 Storage: (to be explored)

Problem: All of the Opensearch nodes use legacy GP2 storage. This costs bit more and only offers thru put based on the volume. i.e using lower disk space might lead to lower thru put, this might be compensated by using higher spec instance sizes.

Solution: Use GP3 storage with varing thru put based on the domain. Initially, achieving a right thru put might be a challenge, but we will gain in long run.

1. AWS Opensearch Warm/cold storage: (to be explored)

Problem: Today all of the storage in Opensearch is Hot storage. This means we are paying more for storage whether we use it or not.

Solution: The application can index the data by warm, cold and there by offloading the cold storage to s3 bucket there by saving cost on hot storage of EBS.

1. Lamba tuning: (to be poc'd) Lambda tuning allows to fine tune the memory parameter by doing performance testing various parameters. This gives a picuture of sweet sport configuration based on various configuration. We could try out most used lambda's using AWS proposed Lambda tuning process and see if we can get savings out of it.