Show notes
Andy Oxfeld, Engineering Manager at Humble Bundle, tells your hosts Francesc and Mark all the details about how Google Cloud Platform powers Humble Bundle.
About AndyAndy Oxfeld manages the engineering department at Humble Bundle. He’s been with the company for four years and worked on almost every part of our codebase at some point. Prior to being at Humble Bundle, he worked at Electronic Arts for four years on games such as Warhammer Online: Age of Reckoning.
Cool thing of the week- An independent organization just ranked Google as the best cloud, beating Amazon news
- Humble Bundle has raised over $70 Million for charity.
- App Engine to handle spiky traffic.
- Codebase written in Python.
- Google Cloud Datastore to store data.
- Task Queues to be able to retry tasks (payments) and to schedule events.
- Memcache to reduce latency and increase throughput.
- Dedicated Memcache for heavy users and more predictable performance.
- Best Practices for App Engine Memcache article.
- Datastore Backups into a Cloud Storage bucket.
- BigQuery to analyze huge datasets in seconds.
- Using only managed services removes their need for DevOps team.
- Google Cloud Storage to store static images.
- Stackdriver for monitoring and alerts.
- PubNub for real time communication - we also discuss Channels API and Firebase.
- Join Humble Bundle and work with Google Cloud Platform!
Kubernetes 1.2 doesn’t have a concept of schedule tasks, how do you handle them?
- Kubernetes jobs docs.
- Cron jobs on Compute Engine.
- Scheduled tasks with Cron on App Engine docs.