AWS CloudWatch Metrics

This article describes how to get started with AWS CloudWatch metrics integration for Opvizor/Cloud
Written by Kristaps
Updated 10 months ago

Overall Metrics Flow


1. Setup Telegraf service with the provided telegraf.conf configuration from Opvizor/Cloud Integrations page.

Make sure to adjust the access_key for secure authentication from AWS Firehose.

Telegraf can be retrieved and installed with the following command lines

wget sudo dpkg -i telegraf_1.26.3-1_amd64.deb_amd64.deb
sudo cp ./telegraf.conf /etc/telegraf/telegraf.conf
sudo systemctl enable telegraf
sudo systemctl restart telegraf

Watch our screen capture to ease your setup.

Note that AWS Kinesis Firehose data stream requires secure endpoint as destination therefore a valid TLS certificate must be set for the telegraf service.

Additionally it is not possible to point the AWS Kinesis Firehose to send metrics to a destination endpoint with port other than 443. To overcome this you can setup a port-forwarding using for example socat and enable it as a system service

Description=Telegraf 443 to 2443 port forwarder


ExecStart=socat TCP-LISTEN:443,fork,reuseaddr TCP:


Copy/Paste the above service manifest to /etc/systemd/system/telegraf-forward.service

Then enable and start the service by executing the following command lines

sudo systemctl enable telegraf-forward.service
sudo systemctl restart telegraf-forward.service

2. Setup AWS Kinesis Firehose

Set the source and destination type for the data pipe

Adjust the destination settings

Additionally you might want to setup the backup of your metrics to s3

3. Enable CloudWatch Metric Streams

We recommend enabling streams for all available metrics however it is always up to you to choose what metric namespaces are selected.

Finally choose the AWS Firehose data pipe as destination and set format of metrics to JSON.

Browsing the Metrics

Once your are done with the configuration allow it to run for 5 - 10 minutes. Then you can start to look at the AWS dashboards.

AWS Lambda




Did this answer your question?