Very simple CloudWatch CLI for Zabbix/Sensu/Mackerel/etc.
Download binary from releases.
Put it into your $PATH.
View it first.
$ metrin --help
Use environmental variables. AWS_REGION, AWS_PROFILE, AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY
Act as Nagios/Sensu plugin style.
Example: Check an EC2 instance's CPU, critical if > 50%
$ metrin --namespace AWS/EC2 --metric-name CPUUtilization --statistic Average --dimension InstanceId:i-1234abcd1234abcde check --critical-gt 50
CloudWatch OK, got `22.534000` (Percent)
Params: {
Dimensions: [{
Name: "InstanceId",
Value: "i-1234abcd1234abcde"
}],
EndTime: 2016-12-04 23:24:55 +0900 JST,
MetricName: "CPUUtilization",
Namespace: "AWS/EC2",
Period: 60,
StartTime: 2016-12-04 23:19:55 +0900 JST,
Statistics: ["Average"]
}
Run command below to view full option list.
$ metrin check --help
Print metric statistic value with given format (--template).
Example: Metric path, value, timestamp, tab separated (Default template)
$ metrin --namespace AWS/EC2 --metric-name CPUUtilization --statistic Average --start-time -900 --period 300 --dimension InstanceId:i-abcd1234abcd12345 print
CloudWatch.InstanceId.i-abcd1234abcd12345.CPUUtilization.Average 1.566 1480861800
CloudWatch.InstanceId.i-abcd1234abcd12345.CPUUtilization.Average 1.536 1480862100
Example: Single value only output (for Zabbix, etc.)
$ metrin --namespace AWS/EC2 --metric-name CPUUtilization --statistic Average --start-time -900 --period 300 --dimension InstanceId:i-abcd1234abcd12345 print --template '{{.Datapoint.Average}}' --last-value-only
1.536
Run command below to view full option list.
$ metrin print --help
This feature is using text/template package.
Additional functions are unixtime (Convert .Datapoint.Timestamp to UNIX timestamp), deref (Convert *float64 to float64, combine usage with printf), getvalue (Get statistic value dynamically).
See print.go.
For troubleshooting.
Adjust --start-time, --end-time, --period.
--dimension dim1key:dim1value --dimension dim2key:dim2value