The Taming of the Queue
Talkat Cloud Foundry Day North America 2024 · on Wednesday, 2024-05-15· at 13:45· Education Center, 2nd Floor·
summary
Only four people in the room have used the autoscaler, yet the case shows a working system that auto-scales background workers by queue depth with the Cloud Foundry App Autoscaler driven by a rabbitmq metrics emitter packaged as a BOSH release. It monitors RabbitMQ message counts and adjusts app instances, removes manual “tap on the shoulder” scaling, survives aggressive replayed loads, and delivers smoother performance, better resource utilization and improved customer-facing reliability.
AI-written summary · about
chapters (16)
- 0:00Introduction and speaker background
- 1:02English comprehension test
- 1:44Benefits of cloud-based solutions
- 2:50Cloud Foundry App Autoscaler overview
- 3:54Key autoscaling metrics explained
- 4:58Custom metrics and their importance
- 6:46Examples of custom metrics
- 8:56Case study introduction
- 9:17Client problem and context
- 10:45Solution design for autoscaling
- 11:52RabbitMQ metrics emitter setup
- 12:34Testing and fine-tuning thresholds
- 13:17Outcomes and benefits achieved
- 15:04Technical architecture overview
- 17:32Reflections and future potential
- 18:15Conclusion and final thoughts
related talks
29mCustom App Autoscaler Using Cloud Controller API - Christopher Decelles
35mAuto-Scaling, Watch My Instances Grow and Shrink
38m50 Shades of K8s Autoscaling - Mario Leander Reimer
01h 29mBatch Processing and Integration on Cloud Foundry
22mIntelligent Autoscaling in Enterprise...
28mAchieving Scale to Zero and Back Again with Cloud Foundry
33mA Tale of Two System Crashes: Innovation and Accidental Chaos
39mManaged CF-as-a-Service: An Operator's Cookbook