Skip to content

Gaming · Kubernetes & AWS migration

MavenHut: stabilizing a gaming platform with Kubernetes.

CloudHero helped MavenHut move from Docker Swarm on DigitalOcean to Kubernetes and AWS. The two-month migration covered applications, databases, deployment pipelines and monitoring.

By Andrei Manea · Published
Updated

Migration delivery
2 months
Daily active users reported in the original study
500,000
MavenHut office with development desks and monitoring screens
MavenHut’s development workspace, pictured in the original case study.

Why MavenHut moved from Docker Swarm

MavenHut develops social games. Its platform ran on DigitalOcean, using Docker Swarm and custom scripts to orchestrate containers. Managing stateful applications was difficult, scaling required additional work, and developers needed assistance to deploy new services.

The team wanted a more stable platform and a clearer way to deploy and monitor applications. CloudHero worked with MavenHut to plan a move to Kubernetes and AWS, covering the application environment and the data services it depended on.

What CloudHero migrated in two months

CloudHero’s Andrei Manea, Andrei Petrescu and George Drăghici worked with MavenHut’s Andrei Văduva on the migration plan. The implementation included:

  1. Applications and search. Move the applications running on Docker Swarm and the Elasticsearch data to Kubernetes.
  2. Deployment pipelines. Create Jenkins pipelines for the Kubernetes environment.
  3. Data services. Move Redis to Amazon ElastiCache, and migrate MySQL and MongoDB data.
  4. Monitoring. Set up Prometheus and Grafana to monitor the platform.

These changes brought application deployment, data migration and operational visibility into the same delivery plan. The work extended beyond replacing the container orchestrator.

Platform stability at 500,000 daily active users

The original case study reported that the migrated platform ran smoothly at 500,000 daily active users. Its PagerDuty reports show a downward trend in recorded incidents across the later reporting period.

PagerDuty monthly incident report from April 2017 to April 2018, with a peak in October 2017
Earlier reporting period: 1 April 2017 to 1 April 2018. Vertical scale: 0–400 incidents.
PagerDuty monthly incident report from April 2018 to April 2019, showing declining recorded incidents
Later reporting period: 1 April 2018 to 1 April 2019. Vertical scale: 0–250 incidents.

These historical PagerDuty screenshots were published with the original case study. The charts use different vertical scales and do not provide exact monthly counts. They show the reported incident trend, rather than a measured percentage reduction or a current service-level guarantee.

Your Kubernetes project

Plan the applications, data and operations together.

Discuss your current platform, deployment constraints and migration requirements with CloudHero.

Contact CloudHero

Tell us what you're working on

Our sales team replies by email within 1 business day.

Topic

We use your details only to reply. Privacy policy