Server monitoring also uses Prometheus and Grafana. Can monitor server system load, CPU usage, network traffic, disk usage, disk read and write speed, IO time consumption, network information.
Install and use
- Install and start node_exporter
- Configuration modification prometheus.yml
- Download Grafana's Node Exporter dashboard
- Import dashboard
It is recommended to understand the principle of Prometheus briefly
Prometheus is CNCF's second largest open source project and is a system and service monitoring system. The first project of the CNCF Foundation is the popular K8s, which collects indicators from configured targets at a given time interval, evaluates rules, displays the results, and supports alarm triggering. At present, more and more projects (such as Kubernetes and etcd) have added support for Prometheus, which reflects its value from the side.
- Prometheus defines the indicator data standard. Data
can be collected by pull, and it supports static collection (configuration file) and dynamic collection (dynamic discovery and dynamic monitoring by means of Zookeeper, Consul, Kubernetes, etc.)
- Prometheus's own time series database
Prometheus Server can quickly analyze and calculate historical data, and it can also be connected to third-party time series databases.
- Prometheus provides PromQL query mode and HTTP query mode
It can easily access various UI systems (Grafana, self-developed test platform) for information display. 4. Prometheus' Alertmanager provides monitoring alarms.
According to the set rules, Prometheus synchronizes alarm information through push operations (supports access to Pageduty, email, and Dingding Wait)
Prometheus is high-energy, fun, recommended, and recommended for learning
Public number: Avenue test