SNMP with Grafana
Grafana Web Interface2/3
-
Add
InfluxDBas a data source in Grafana using the credentials and bucket name specified in your.envfile.- Query language:
Flux - URL:
http://influxdb:8086 - Organization:
buildit - Token:
buildit-monitoring-token - Default Bucket:
network
Click
Save & Testto verify the connection to InfluxDB.
- Query language:
-
Create a dashboard to visualize the SNMP metrics collected by Telegraf.
Grafana Web Interface3/3
-
Add panels to the dashboard to display different types of data, such as network traffic, CPU usage, and memory utilization.
The full documentation for creating dashboards and panels in Grafana can be found in the Grafana documentation.
The queries below are some examples that retrieve the SNMP metrics collected by Telegraf from InfluxDB for a panel. Some metrics can be displayed directly, while others need to be calculated or converted before they are useful for visualization.
System Uptime
Section titled “System Uptime”System Uptime from(bucket: "network")|> range(start: v.timeRangeStart, stop: v.timeRangeStop)|> filter(fn: (r) => r._measurement == "snmp")|> filter(fn: (r) => r._field == "uptime")|> last()|> map(fn: (r) => ({ r with _value: float(v: r._value) / 100.0 }))CPU Utilization
Section titled “CPU Utilization”hrProcessorLoad(and consequently thecpuandload_percentdata we are importing to InfluxDB) reports utilization separately for each logical processor. This query retrieves the latest reading for each processor and calculates their mean to produce an overall CPU utilization percentage.Current CPU Utilization from(bucket: "network")|> range(start: v.timeRangeStart, stop: v.timeRangeStop)|> filter(fn: (r) => r._measurement == "cpu")|> filter(fn: (r) => r._field == "load_percent")|> last()|> group()|> mean()Memory Utilization
Section titled “Memory Utilization”Linux uses otherwise-unused memory for caching, so the amount of memory marked as “used” does not necessarily represent memory pressure. Instead, we’ll calculate utilization using available memory:
Memory Utilization = (1 − Available Memory / Total Physical Memory) × 100
Memory Utilization total = from(bucket: "network")|> range(start: v.timeRangeStart, stop: v.timeRangeStop)|> filter(fn: (r) => r._measurement == "storage")|> filter(fn: (r) => r.storage == "Physical memory")|> filter(fn: (r) => r._field == "size")|> last()|> findRecord(fn: (key) => true, idx: 0)available = from(bucket: "network")|> range(start: v.timeRangeStart, stop: v.timeRangeStop)|> filter(fn: (r) => r._measurement == "storage")|> filter(fn: (r) => r.storage == "Available memory")|> filter(fn: (r) => r._field == "size")|> last()|> findRecord(fn: (key) => true, idx: 0)from(bucket: "network")|> range(start: v.timeRangeStart, stop: v.timeRangeStop)|> filter(fn: (r) => r._measurement == "storage")|> filter(fn: (r) => r.storage == "Physical memory")|> filter(fn: (r) => r._field == "size")|> last()|> map(fn: (r) => ({r with_value: (1.0 - (float(v: available._value) / float(v: total._value))) * 100.0}))Disk Utilization
Section titled “Disk Utilization”Disk Utilization = Used Space / Total Space × 100
Disk Utilization size = from(bucket: "network")|> range(start: v.timeRangeStart, stop: v.timeRangeStop)|> filter(fn: (r) => r._measurement == "storage")|> filter(fn: (r) => r.storage == "/")|> filter(fn: (r) => r._field == "size")|> last()|> findRecord(fn: (key) => true, idx: 0)from(bucket: "network")|> range(start: v.timeRangeStart, stop: v.timeRangeStop)|> filter(fn: (r) => r._measurement == "storage")|> filter(fn: (r) => r.storage == "/")|> filter(fn: (r) => r._field == "used")|> last()|> map(fn: (r) => ({r with_value: float(v: r._value) / float(v: size._value) * 100.0}))Network Interface Status
Section titled “Network Interface Status”Status of Network Interface eth0 from(bucket: "network")|> range(start: v.timeRangeStart, stop: v.timeRangeStop)|> filter(fn: (r) => r._measurement == "interface")|> filter(fn: (r) => r.interface == "eth0")|> filter(fn: (r) => r._field == "status")|> last()Network Interface Speed
Section titled “Network Interface Speed”eth0 Speed Capacity from(bucket: "network")|> range(start: v.timeRangeStart, stop: v.timeRangeStop)|> filter(fn: (r) => r._measurement == "interface")|> filter(fn: (r) => r.interface == "eth0")|> filter(fn: (r) => r._field == "speed_mbps")|> last()Network Traffic
Section titled “Network Traffic”Earlier, we calculated a traffic rate by finding the difference between two SNMP counter readings over time. The derivative() function performs that calculation automatically. Because ifHCInOctets and ifHCOutOctets count bytes, multiplying the resulting bytes-per-second value by 8 converts it to bits per second.
eth0 Inbound Traffic from(bucket: "network")|> range(start: v.timeRangeStart, stop: v.timeRangeStop)|> filter(fn: (r) => r._measurement == "interface")|> filter(fn: (r) => r.interface == "eth0")|> filter(fn: (r) => r._field == "in_octets")|> derivative(unit: 1s, nonNegative: true)|> map(fn: (r) => ({r with_value: float(v: r._value) * 8.0}))|> set(key: "_field", value: "Inbound")|> keep(columns: ["_time", "_value", "_field"])eth0 Outbound Traffic from(bucket: "network")|> range(start: v.timeRangeStart, stop: v.timeRangeStop)|> filter(fn: (r) => r._measurement == "interface")|> filter(fn: (r) => r.interface == "eth0")|> filter(fn: (r) => r._field == "out_octets")|> derivative(unit: 1s, nonNegative: true)|> map(fn: (r) => ({r with_value: float(v: r._value) * 8.0}))|> set(key: "_field", value: "Outbound")|> keep(columns: ["_time", "_value", "_field"])Network Errors and Discards
Section titled “Network Errors and Discards”eth0 Errors and Discards from(bucket: "network")|> range(start: v.timeRangeStart, stop: v.timeRangeStop)|> filter(fn: (r) => r._measurement == "interface")|> filter(fn: (r) => r.interface == "eth0")|> filter(fn: (r) =>r._field == "in_errors" orr._field == "out_errors" orr._field == "in_discards" orr._field == "out_discards")|> derivative(unit: 1s, nonNegative: true)|> keep(columns: ["_time", "_value", "_field"])