oslo.metrics -- Monitoring OpenStack RPC Calls - Ching Kuo

Understanding the scale limitation and identifying the bottleneck are essential to provide a reliable large-scale cloud. The scale of the OpenStack cloud in LINE exceeded 2000+ hypervisors and started facing scalability and performance issues related to RPC. We are monitoring the RabbitMQ, however, even though RabbitMQ is working well, we can not guarantee RPC is working perfectly from OpenStack perspective. To monitor the RPC in detail, we developed oslo.metrics which exposes the internal metrics of oslo libraries and integrated it with oslo.messaging. This talk will give an introduction to oslo.metrics, describing how we collect the internal metrics and what kind of metrics we are collecting. We will also demonstrate how to try oslo.metrics using devstack and the current progress of development upstream.

Видео oslo.metrics -- Monitoring OpenStack RPC Calls - Ching Kuo автора Кодовый революционер с Python