Node_exporter monitoring indicators

1. Node_exporter monitors agent survival

PromQL statement:

up{job="捷顺监控"} == 0

Description: Check whether the Node_exporter monitoring agent is alive, or whether the Node_exporter service is down

2. Node_exporter monitors CPU utilization

PromQL statement:

ceil(100 - sum(increase(node_cpu_seconds_total{job="捷顺监控",mode="idle"}[5m]))  by(instance) / sum(increase(node_cpu_seconds_total{job="捷顺监控"}[5m]))  by(instance)*100) >= 80

Description: Monitor the CPU utilization of the server, if it is greater than 80%, it will alarm

3. Node_exporter monitors disk utilization

PromQL statement:

round((1 - (node_filesystem_avail_bytes{fstype=~"ext3|ext4|xfs|nfs",job="捷顺监控"} / node_filesystem_size_bytes{fstype=~"ext3|ext4|xfs|nfs",job="捷顺监控"})) * 100)  >= 80

Description: Monitor the disk utilization of the server, and alarm if it is greater than 80%

4. Node_exporter monitors memory utilization

PromQL statement:

ceil((1 - (node_memory_MemAvailable_bytes{job="捷顺监控"} / (node_memory_MemTotal_bytes{job="捷顺监控"})))* 100 ) >= 80

Description: Monitor the memory utilization of the server, and alarm if it is greater than 80%

5. Node_exporter monitors server load

PromQL statement:

node_load1{job="捷顺监控"} >=50

Description: Monitor the server load of the server, if it is greater than 50, it will alarm

6. Node_exporter monitors server file handle usage

PromQL statement:

node_filefd_allocated{job="捷顺监控"} >=50000

Description: Monitor the server's file handle usage, if it is greater than 50000, it will alarm

7. Node_exporter monitors the number of TCP connections waiting to be closed by the server

PromQL statement:

node_sockstat_TCP_tw{job="捷顺监控"} >=5000

Description: Monitor the number of TCP connections waiting to be closed by the server. If it is greater than 5000, it will alarm

8. Node_exporter monitors server entrance traffic

PromQL statement:

round((sum
  by(instance) (irate(node_network_receive_bytes_total{device!~"tap.*|veth.*|br.*|docker.*|virbr*|lo*",job="捷顺监控"}[5m])))
  / 1024 / 1024) > 50

Description: Monitor the server inlet flow, if the inlet flow is greater than 50M, it will alarm

9. Node_exporter monitors server export traffic

PromQL statement:

round((sum
  by(instance) (irate(node_network_transmit_bytes_total{device!~"tap.*|veth.*|br.*|docker.*|virbr*|lo*",job="捷顺监控"}[5m])))
  / 1024 / 1024) > 50

Description: Monitor the server's export traffic, if the export traffic is greater than 50M, it will alarm

Guess you like

Origin blog.csdn.net/qq_31555951/article/details/110666480