Built by Anrotex

Reliable operations. Checked releases. Clear recovery.

Cloud Operations brings an authenticated requests API, repeatable AWS infrastructure and a tested release procedure together in one engineering project.

The problem we set out to solve

A retry should not create a duplicate operation. A deployment should prove that the application works. And a failed release should leave the team with a clear recovery action. We built Cloud Operations to connect those decisions in a small, inspectable system.

The API records operational requests; it does not execute infrastructure changes submitted by callers.

IAM-authenticated requests pass through API Gateway and the Lambda live alias to DynamoDB. GitHub Actions uses an OIDC release role to check candidates and promote or restore the alias. CloudWatch feeds SNS alerts.
The AWS design and release path. Open the diagram to inspect it at full size.

Requests that survive retries

A Python API validates requests, checks ownership and uses atomic DynamoDB writes. Repeating the same request with the same key returns the existing record instead of creating a duplicate.

Infrastructure your team can review

Terraform defines API Gateway, Lambda, DynamoDB, scoped IAM permissions, logs, alarms and budget notifications. The infrastructure and application releases have separate review steps.

A checked path to release

The release script probes a published candidate before moving the live alias. If the subsequent API Gateway checks fail, it requests restoration of the previous version, guarded against concurrent alias changes.

Recovery instructions alongside the code

The project includes a local backup and restore exercise, release failure tests and an operating runbook. The workflow uses temporary AWS credentials through GitHub OIDC.

Retry and recovery checks

A local exercise sent 100 requests using the same retry key: one record was created and 99 attempts returned it. A separate backup exercise restored 25 records and confirmed that their contents matched.

These checks demonstrate application behavior under the tested conditions. Live service limits, permissions and recovery timing need their own AWS acceptance tests.

Apply the approach to your workload

Start with an existing AWS workload or one application's release pipeline. We agree the scope, required access and acceptance checks before work starts.

Need broader infrastructure work? Explore our DevOps and Terraform services.