This repository was archived by the owner on Mar 10, 2018. It is now read-only.
Detecting system load bottleneck and trace it to the process/container #119
Description
Task for SRE (service reliability engineering) specialists. We've got a single node (hardware machine) running around 24 containers with various interconnected services. At some point API calls stop to respond with timeout. We need to monitor server load, detect bottleneck in system resources - SSD, HDD, setup alerting and automatically detect who is waiting for resoures (process with long disk write queue) and who is causing resource depletion.
- propose strategy for resolving the issue
- get access to hardware
- setup monitoring services
- show demo for development team
- profit !!!