2

I'm trying to implement an alarm(in Cloudformation) for disk space free using metrics from the Cloudwatch agent and I'm having issues with devices shuffling DeviceID.

I encountered this earlier when working with fstab, EC2 instances attach EBS volumes in a seemingly random order. If I have two EBS volumes, one for OS and one for data, they could randomly swap between nvme0n1/nvme1n1. I was able to solve this using partition UUIDs.

However I see no way to work around this when creating alarms. Alarms must be created with metrics that specify the DeviceID(nvme0n1) and the mount point(/var/data). These can swap at any given boot. See https://imgur.com/a/XJ62zN8 You can see there's four data points for effectively two points of information as the devices change when the server reboots.

Is there some way to make Cloudwatch Agent report without the Device or to write the alarm to apply to any Device metric?

1

After doing some digging in the documentation the solution is to drop the device by using a boolean in the config json file.

https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/CloudWatch-Agent-Configuration-File-Details.html

The resulting block will look like this.

            "disk": {
            "measurement": [
                "used_percent"
            ],
            "metrics_collection_interval": 60,
            "resources": [
                "/",
                "/var/data"
            ],
            "drop_device": true
        },
| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.