Case Study : Improving Resilience of Applications in Telco Environme...
Talkat KubeCon + CloudNativeCon North America 2021 · on Wednesday, 2021-10-13· at 11:00· Room 408 AB + Online·
summary
Orange runs OpenStack on Kubernetes at telco scale and uses Litmus to run repeatable chaos workflows with probes and CrossTesting for custom validations; Litmus claims ~1,000 installs/day and ships Litmus 2.0 as the orchestration and probe engine. A Vault experiment deleting pods in parallel unseals and passes post‑chaos probes (Vault pod parallel delete -> recovered), while serial deletes intentionally fail the continuous URL probe and report serial delete -> 50% probe success. A Nova scheduler pod‑delete test creates VMs during chaos and shows no packet loss, indicating resilience.
AI-written summary · about
chapters (23)
- 0:00About Speakers
- 0:27The goal
- 0:44In this session ...
- 1:29Resilience challenge in cloud native
- 2:35Solution to resilience challenge
- 4:43Introduction to Litmus Chaos
- 5:51About Litmus Chaos
- 6:33Where to use LitmusChaos?
- 8:26Getting started with Litmus
- 9:11Environment and scenario
- 10:34Complexities of Telco Infra
- 11:29Chaos and Resiliency Testing
- 12:08System Under Test
- 13:06Why Litmus ?
- 13:59Building customized Probes
- 14:49Chaos Testing Flow
- 15:53Use Case
- 16:53Delete vault pod
- 21:22Parallel chaos into Vault
- 25:51Demo 3 - Chaos into Nova Schedule
- 31:09Results
- 32:30Summary
- 32:59Reliability for your DevOps?
related talks
39mPracticing Cloud-Native Resliency Is a Challenge: Decoding It with Chaos...
01h 01mMastering Resilience Probes: Simplify chaos hypot...
30mThe Narrative of Reliability: A Chaos Engineering Case Study with LitmusChaos
26mCanonical chaos through Litmus
29mUnveiling Resilience: iFood's Jenga & Sherlock Holmes Practices in Chaos Engineering
57mChaos Engineering Primer
29mBoF Deep Dive: Chaos Engineering
25mResilience testing in modernization projects