Anbox Cloud allows you to collect various metrics from a deployment. This includes various metrics about the hardware and how it’s being used but also about Anbox Cloud itself, like how many containers are currently running etc.
The following table shows an overview of the metrics available through Prometheus:
| Name | Description |
|---|---|
nodes |
The number of nodes in the cluster |
applications |
The number of applications in the cluster |
containers |
The number of containers in the cluster |
container_boot_time |
The number of seconds it takes to boot a container |
container_per_application |
The number of containers per application |
containers_per_status |
The number of containers per container status |
available_cpu |
The total number of CPUs available in a node of the cluster |
used_cpu |
The number of CPUs used in a node of the cluster |
available_memory |
The total memory available in a node of the cluster |
used_memory |
The memory used in a node of the cluster |
Requirements
In order to collect various metrics from the LXD machines Anbox Cloud uses Telegraf. For visualization purposes Grafana is being used.
To make the deployment of all necessary components easy and straight forward a Juju bundle called cs:~anbox-charmers/anbox-cloud-metrics is available in the Juju charm store.
By default the bundle deploys Prometheus and Grafana into a new machine and will make access to Grafana available through a HTTPS secured proxy powered by Apache.
Tip: When you’re structuring your production deployment you may want to follow our instructions for a customized deployment and directly include any relevant components in your own deployment bundle.
More details about using the deployed metrics collection stack can be found here.
Deployment
You can add the metrics services to an existing Anbox Cloud deployment by executing:
$ juju deploy cs:~anbox-charmers/anbox-cloud-metrics
If you’re using the manual Juju provider as described in the custom install guide you can map the machine defined in the bundle to an existing one on your bundle:
$ juju deploy cs:~anbox-charmers/anbox-cloud-metrics --map-machines 0=2
As in previous deployment steps, you can check the status of the deployment with:
$ watch -c juju status --color --relations=true
When the deployment finished, the status should look like this.
Model Controller Cloud/Region Version SLA Timestamp
default anbox-cloud aws/us-east-2 2.5.0 unsupported 16:03:32+08:00
App Version Status Scale Charm Store Rev OS Notes
ams active 1 ams local 0 ubuntu
ams-node-controller active 1 ams-node-controller local 0 ubuntu exposed
easyrsa 3.0.1 active 1 easyrsa jujucharms 195 ubuntu
etcd 3.2.10 active 1 etcd jujucharms 338 ubuntu
grafana active 1 grafana jujucharms 22 ubuntu
grafana-proxy unknown 1 apache2 jujucharms 26 ubuntu exposed
lxd active 1 lxd local 0 ubuntu
prometheus active 1 prometheus2 jujucharms 8 ubuntu
telegraf active 1 telegraf jujucharms 27 ubuntu
Unit Workload Agent Machine Public address Ports Message
ams/0* active idle 0 18.224.58.129 8443/tcp,9103/tcp
easyrsa/0* active idle 0 18.224.58.129 Certificate Authority connected.
etcd/0* active idle 0 18.224.58.129 2379/tcp Healthy with 1 known peer
grafana-proxy/0* unknown idle 2 18.216.182.177 80/tcp,443/tcp
grafana/0* active executing 2 18.216.182.177 3000/tcp Started grafana-server
lxd/0* active idle 1 52.14.125.184 8095/tcp,8443/tcp
ams-node-controller/0* active idle 52.14.125.184 10000-11000/tcp
telegraf/0* active idle 52.14.125.184 9103/tcp Monitoring lxd/0
prometheus/0* active executing 2 18.216.182.177 9090/tcp,12321/tcp Ready
Machine State DNS Inst id Series AZ Message
0 started 18.224.58.129 i-0368b9501551aed87 bionic us-east-2a running
1 started 52.14.125.184 i-019d854aed4ad262e bionic us-east-2b running
2 started 18.216.182.177 i-01f7943b88b44e212 bionic us-east-2c running
Relation provider Requirer Interface Type Message
ams:prometheus prometheus:scrape prometheus regular
easyrsa:client etcd:certificates tls-certificates regular
etcd:cluster etcd:cluster etcd peer
etcd:db ams:etcd etcd regular
grafana:website grafana-proxy:reverseproxy http regular
lxd:api ams-node-controller:lxd rest subordinate
lxd:api ams:lxd rest regular
lxd:juju-info telegraf:juju-info juju-info subordinate
prometheus:grafana-source grafana:grafana-source grafana-source regular
telegraf:prometheus-client prometheus:target http regular