Orcha 🐳: Massive Parallel Processing (MPP) and API design
summary
Orcha is a Python-based Massively Parallel Processing orchestrator that parallelizes CI by using only Python multiprocessing plus a plugin model (docker, vagrant, Robot Framework). It targets huge test-suite speedups (examples: smoke from 2h to 20min, eliminating 15h weekly runs, running ~200 nightly jobs in parallel) and provides an API with managers, messages, petitions, start/finish/cancel hooks, background sync, composition instead of inheritance, and plugin-controlled reservation/release of real resources (RAM, CPU, devices) for stable daily operation.
AI-written summary · about
chapters (7)
- 0:00Start
- 8:36Manager
- 19:10Inheritance vs Composition: Program
- 24:52Speed
- 25:14Flexibility
- 25:37Stability
- 25:55Future
related talks
32mSecOps Automation And Orchestration
18mCluster Management: Heterogeneous, Lightweight, Safe. Pick Three
25mOne Engine To Rule Them All: Unifying Cloud Workloads With Argo Workflows
34mWorkflow Orchestration Embedded in your DevOps Tool Chain
16mToward Orchestration of Complex Networking Experiments
45mOrchestrating Tasks across Hybrid Environments using ...
21mOrchestrating multi-cloud data pipelines with
33mUnleashing the Gorillacorn - DevOps in Very Large Enterprises