Open Networking Summit Europe 2018

This presentation shows that Prometheus can bridge the gap between the Telco NFV world and the container world, and is applicable for any platform that requires monitoring. The transition to containers in Telco is happening, but not fast enough.

The session will present a completed PoC where a Prometheus Operator was used to monitor an NFVI+VIM targeted for 1000 nodes. 

Data collection was provided by the OPNFV Barometer project's collectd container, data transport by AMQP1.0, and metric storage by Prometheus. System events and logs were stored into ElasticSearch. Alert Manager was used to send alerts via SNMP in response to metric conditions monitored by Prometheus.

A test framework that simulated 1000 nodes (200 metrics/node) were developed and was used to evaluate the performance of Prometheus for storage and alerting.