Warning
You are currently viewing v"2.4" of the documentation and it is not the latest. For the most recent documentation, kindly click here.
The metrics exposed by KEDA Metrics Server can be queried directly using kubectl
:
kubectl get --raw "/apis/external.metrics.k8s.io/v1beta1"
This will return a json with the list of metrics exposed by KEDA:
{
"kind": "APIResourceList",
"apiVersion": "v1",
"groupVersion": "external.metrics.k8s.io/v1beta1",
"resources": [
{
"name": "s0-rabbitmq-queueName",
"singularName": "",
"namespaced": true,
"kind": "ExternalMetricValueList",
"verbs": [
"get"
]
},
{
"name": "s1-rabbitmq-queueName2",
....
}
]
}
You can also query for the value of a specific metric using kubectl
:
kubectl get --raw "/apis/external.metrics.k8s.io/v1beta1/namespaces/YOUR_NAMESPACE/YOUR_METRIC_NAME"
At this point, you should take in consideration that KEDA metrics are namespaced, this means that you have to specify the namespace where the ScaledObject
is placed inside.
For example, if you want to get the value of the metric named s1-rabbitmq-queueName2
in namespace sample-ns
, the query will be like this:
kubectl get --raw "/apis/external.metrics.k8s.io/v1beta1/namespaces/sample-ns/s1-rabbitmq-queueName2"
And it will show a json like this:
{
"kind": "ExternalMetricValueList",
"apiVersion": "external.metrics.k8s.io/v1beta1",
"metadata": {},
"items": [
{
"metricName": "s1-rabbitmq-queueName2",
"metricLabels": null,
"timestamp": "2021-10-20T10:48:17Z",
"value": "0"
}
]
}
Note: There are 2 exceptions in querying metrics and those are
cpu
andmemory
scalers. When KEDA creates the HPA object, it uses standardcpu
andmemory
metrics from the Kubernetes Metrics Server. If you want to query these 2 specific values, you should do it using/apis/metrics.k8s.io/v1beta1
instead of/apis/external.metrics.k8s.io/v1beta1
.
During its work, KEDA updates each ScaledObject with some relevant information which it needs to work. Part of that information is metric names generated from the triggers inside the own ScaledObject.
You can recover the metric names from a ScaledObject using kubectl
:
kubectl get scaledobject SCALEDOBJECT_NAME -n NAMESPACE -o jsonpath={.status.externalMetricNames}
KEDA will try to select the proper ScaledObject
for your metric and there should only be one. In case of having multiple ScaledObject
s in the same namespace with the same metric name, an error like this will be thrown:
Error from server: exactly one ScaledObject should match label
In this case, you should add in the query string the labelSelector
to match the proper ScaledObject
(in url format). The needed selector is scaledobject.keda.sh/name: {ScaledObjectName}
. You can achieve this as following:
kubectl get scaledobject SCALEDOBJECT_NAME -n NAMESPACE -o jsonpath={.metadata.labels}
Once you have the selector, you have to add it to your query string as following:
kubectl get --raw "/apis/external.metrics.k8s.io/v1beta1/namespaces/sample-ns/s1-rabbitmq-queueName2?labelSelector=scaledobject.keda.sh%2Fname%3D{ScaledObjectName}"