Skip to main content

Cost optimization for Live Stream Marketplace

Client was not happy with their AWS bills and asked us to make it lean as possible without any degradation of their services performance and disaster recovery.

It's a product to make video presentation video streaming providing stable  and high quality solution. So, it's classical high load with 24/7 requirements and quite low cost for client. That's why cost of cloud infrastructure is quite important for business success. There is a few huge services it integrates to, so client has to provide streaming for with strict SLA.

Solution is based on Ruby, RoR, PostgreSQL and MongoDB,  VMs and dedicated services with bunch of other services and dozens of 3rd party integrations. It took just a couple of hours speaking with System Architect and just our efforts - and we renew to company their infrastructure and deployment documentation (it was outdated) and point on some issues that can be easily fixed.

All the project took a month, but first significant results client got in a 2 weeks. 

These actions helped to reduce AWS monthly cost by 47%.

What client get as a benefit, except it?

We depliy additional pre-prod environment dedicated to demo and integration testing purpose. And we make some improvements in monitoring service uptime and performance.

We made inventory of AWS services used. There were EC2, S3, CloudFront, Route53.

We installed monitoring tools and started to gather metrics from all EC2 instances.

We analyzed S3 content and found content that could be deleted.

We analyzed metrics from the monitoring and defined EC2 instances that could have other instance type.

We analyzed solution architecture and rebuilt it by combining some services on the same instance, not separate.

We analyzed and redesign backup strategy including configuration of old snapshots and dumps removal.

Finally we developed and performed load tests to be sure that our changes led to desired performance, not less.

The next steps are development of micro-services architecture, dockerization and deployment to K8s cluster with auto up and down scaling.

DevOps Automation Professional Service
Our team builds, automates, containerizes, migrates, secures, optimizes, makes compliant — and creates harmony