Covering announcements from 9 to 15 November.
What changed: You can now restore an expired or recently deleted domain from the Route 53 console, the CLI or the SDKs. Before, it needed a request to AWS Support.
Who it affects: Anyone whose domains are registered in Route 53.
What to do: A restore usually costs more than a renewal, so prevention is still better: check auto-renew is on and the payment card on the account is current.
What changed: You can lock important records (for example your website and email records) in the hosted zone, so changes are blocked until someone authorised removes the protection.
Who it affects: Any business whose website or email depends on DNS in Route 53.
What to do: Protect the records that would take your website or email offline if someone changed them by mistake.
What changed: Logically air-gapped vaults (locked, isolated backups that can be restored directly) can now move rarely used EBS backups to EBS Snapshots Archive.
Who it affects: Teams that keep EBS backups for months or years.
What to do: If all your backups sit in the same account as production, an isolated vault is worth considering; ransomware and mistakes in that account then cannot delete them.
What changed: A secret can now be replicated to a second AWS account and kept up to date automatically, including after rotation.
Who it affects: Companies with several AWS accounts, or planning a separate recovery account.
What to do: Include your critical secrets in your disaster recovery plan; without them, restored systems often cannot start.
What changed: New R9g instances are available for Aurora (PostgreSQL and MySQL) and RDS for PostgreSQL, MySQL and MariaDB. AWS quotes better performance and price-performance than Graviton4, depending on the engine and workload.
Who it affects: Teams running memory-heavy databases on older instance types.
What to do: Test in a non-production copy first and compare cost and performance before switching production.