Closed danzipie closed 5 years ago
telegraf/linux reports 100% per core. so if there is an 8-core system it can go up to 800% CPU. Azure might be measuring total utilization. Also depends on the averaging window.
I see. The strange thing is that the value is bouncing between 10% (that is the correct value) and more than 100%. I will look again into this and eventually reopen.
I compared Azure monitoring and Telemetry data: they look very different. See CPU usage of the same bootnode (
bootnode3
)Graph on the right generated with:
SELECT mean("usage_percent") FROM "docker_container_cpu" WHERE ("container_name" = 'docker-stack_parity_1' AND "nodetype" = 'proxy') AND $timeFilter GROUP BY time($__interval), "engine_host" fill(linear)