Faults are the primary culprits of breaking the high availability of cloud systems, even leading to costly outages. As the scale and complexity of clouds increase, it becomes extraordinarily difficult to understand, detect and diagnose faults. During outages, engineers record the detailed information of the whole life cycle of faults (i.e., fault occurrence, fault detection, fault identification, and fault mitigation) in the form of post-mortems. In this paper, we conduct a quantitative and qualitative study on 354 public post-mortems collected in three popular large-scale clouds, 97.7% of which spans from 2015 to 2021.
This repository refers to the paper published on ISSRE'22 "Going through the Life Cycle of Faults in Couds: Guidelines on Fault Handling". The paper can be downloaded from 2022-ISSRE-xiaoyun.pdf
Nowadays, many companies make their post-mortems public online. We have collected several data source links as below:
- https://github.com/danluu/post-mortems
- https://ucare.cs.uchicago.edu/projects/cbs/ (Chicago datasets)
- https://dsl.uwaterloo.ca/projects/neat/ (Unaccessible, latest version is required)
- https://docs.google.com/spreadsheets/d/1YLuOTuve_naKwdZBnomzmiQW2XOn0rn-S7VvQ6YCtvc/edit#gid=1
We collected a total of 354 public post-mortems from three popular clouds spanning from 2011 to 2021. We transform unstructured post-mortems as structured ones.
Please cite our ISSRE'22 paper if you find this work is helpful.