AWS service notes rolling out

Amazon Web Services spent April 9, 2026 shipping three small changes that all target the same pain point: production systems usually fail at the handoff, not in the demo. One update touches search and metrics, one touches database cutovers, and one touches object storage cleanup. (aws.amazon.com) The search piece is about observability, which is the practice of watching a live system through logs, metrics, and traces the way a pilot watches gauges in a cockpit. Amazon OpenSearch Service said it now has native integration with Amazon Managed Service for Prometheus, Amazon’s managed metrics store, plus agent tracing features. (aws.amazon.com) Prometheus is built for time-series metrics, which are measurements like request rate or CPU use collected every few seconds and stored as a running timeline. Amazon Managed Service for Prometheus is the hosted version, and AWS says the new integration lets teams consolidate observability without duplicating data across separate tools. (aws.amazon.com 1) (aws.amazon.com 2) AWS had already added a Prometheus sink to Amazon OpenSearch Ingestion in March 2026, so pipelines could send metrics into a Prometheus workspace. The April 9 release pushes that pairing further by tying OpenSearch Service itself to Managed Service for Prometheus and adding tracing for the agents that move telemetry around. (aws.amazon.com 1) (aws.amazon.com 2) The database piece is about blue/green deployment, which is a safety pattern where you run the old production database and a synchronized copy side by side before switching traffic. Amazon Relational Database Service documents blue/green deployments as a separate staging environment that stays in sync with production until switchover. (docs.aws.amazon.com) On April 9, AWS said Amazon Relational Database Service Proxy now supports those blue/green switchovers for single-Region setups. The proxy sits between an app and the database, so instead of every app server needing to discover the new writer after cutover, the proxy detects that the green environment is now production and redirects connections. (aws.amazon.com) That change lands after AWS said on January 20, 2026 that blue/green switchover downtime had been cut to typically five seconds or less for single-Region configurations. Pairing that faster switchover with a proxy is the practical part: the database can flip quickly, and the application can recover quickly too. (aws.amazon.com 1) (aws.amazon.com 2) The storage piece is about lifecycle rules, which are automatic instructions that move old files to cheaper storage or delete them after a set age. Amazon Simple Storage Service says each bucket can have up to 1,000 lifecycle rules, and those rules run automatically once configured. (docs.aws.amazon.com) On April 9, AWS said Amazon Simple Storage Service Lifecycle will now pause expiration and transition actions on objects that failed replication. If a bucket is supposed to copy data to another location and that copy fails because of permissions or replication configuration, Lifecycle will stop short of deleting or archiving the source object underneath it. (aws.amazon.com) That closes a nasty edge case for backup and compliance workflows. Amazon Simple Storage Service documentation already said lifecycle actions are suspended until replication status becomes COMPLETED or FAILED for certain cases, and the new release makes failed replication a clearer stop signal so teams can fix the rule before data is aged out. (docs.aws.amazon.com) (aws.amazon.com) None of these launches is flashy on its own, but they all shave risk off routine operations: watching systems, switching databases, and cleaning up storage. That is usually where cloud platforms win or lose trust, because production outages are often caused by the plumbing around the service, not the service itself. (aws.amazon.com) (aws.amazon.com) (aws.amazon.com)