OK, did a little playing. Here’s what I found.
I created a fresh AmazonLinux 2 EC2 instance and configured the EPEL repository on it (all commands run with root privileges):
amazon-linux-extras install epel -y
At this point, all I should need to do is run “yum install netdata” but it fails out of the box because it requires nodejs 16 and nodejs 16 has a dependency on libuv-1.43.0 which is not available through EPEL or the AmazonLinux 2 repos. The latest available is 1.36.1.
To work around this, you must first get nodejs installed and that installation must be forced to not pay attention to the version of libuv. So install libuv so the wrong (1.36.1) version of the dependency exists and then use rpm --nodeps to force install nodejs:
yum install libuv -y
rpm -Uvh --nodeps $(repoquery --location nodejs)
At this point I installed the other dependencies that seemed to be needed by netdata:
yum install brotli libbsd libretls netcat nodejs-libs openssl11 protobuf protobuf-c snappy -y
and was then able to install netdata. I suspect I could have skipped the previous step and gone straight to the netdata install since nodejs was already installed:
yum install netdata netdata-conf netdata-data -y
I then enabled and started netdata and then proceeded to claim my new node using the --claim-only option of the kickstart script:
systemctl enable netdata
systemctl start netdata
wget -O /tmp/netdata-kickstart.sh https://my-netdata.io/kickstart.sh && sh /tmp/netdata-kickstart.sh --claim-only --claim-token 4xvgbyxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxOD9pME --claim-url https://app.netdata.cloud
This new node was then visible on the netdata cloud interface.
I tested running some docker containers:
amazon-linux-extras install docker
systemctl enable docker
systemctl start docker
docker pull wordpress && docker run --name wordpress -p 443:80 wordpress &
docker pull nginx && docker run --name nginx -p 80:80 nginx &
docker run --name nginx-8081 -p 8081:80 nginx &
But while my running containers were visible on the local port 19999 interface:
They again are not visible by default on the cloud interface:
Can you suggest a config file tweak to make them show up on the cloud interface?
It would be nice if your default settings picked them up for the cloud interface.
-RP