Batch Scheduling Across Multiple Kubernetes Clusters - Mirantis Labs Tech Talks

Explore batch scheduling across multiple Kubernetes clusters in this 33-minute tech talk by Mirantis Labs. Dive into the Armada project, an open-source solution for running batch workloads in multi-cluster Kubernetes environments. Learn about batch scheduling, processing, and real-world applications of open-source tooling. Discover how Armada addresses limitations of existing CNCF projects for batch workflows. Follow along with a live demo showcasing Armada's capabilities in multi-cluster K8s environments. Gain insights into parallel computing, batch schedulers, Kubernetes integration, and open-source community involvement. Understand Armada's features including queue management, RBAC, pod specs, and client support for various programming languages. Explore real-world use cases with Apache Airflow and Jupyter integration. Conclude with valuable CNCF resources related to the Armada project.

Our next installment is set to premiere a week from Tuesday, April 18, 2023 at AM PDT. If you are interested in more upcoming & on demand Tech Talks from the teams at Mirantis Labs, please visit https://www.mirantis.com/labs
- Introducing our presenter
- Some background on G Research & Armada
- Overview of Parallel Compute
- SLURM vs. HTCondor Batch Schedulers
- Then Kubernetes happened...
- Interfacing with multiple K8s clusters
- Batch Processing with Kubernetes
- Kubernetes Batch Processing open source projects
- The Armada solution
- Queue & RBAC
- Pod Specs AKA, Kubernetes constructs
- Clients: DotNet, Python, Go & CLI
- Real world use case: Apache Airflow
- Armada Jupyter integration
- Get involved in the open source K8s community
Live Demo: Armada Batch Scheduling
- CNCF resources related to Armada
- Outro