Memory and cpu resources of a container can be tracked using prometheus. But can we track I/O of a container? Are there any metrices available?
If you are using Docker containers you can check the data with the docker stats command (as P... mentioned in the comment). Here you can find more information about this command.
If you want to check pods cpu/memory usage without installing any third party tool then you can get memory and cpu usage of pod from cgroup.
Go to pod's exec mode kubectl exec pod_name -- /bin/bash
Go to cd /sys/fs/cgroup/cpu for cpu usage run cat cpuacct.usage
Go to cd /sys/fs/cgroup/memory for memory usage run cat memory.usage_in_bytes
For more look at this similar question.
Here you can find another interesting question. You should know, that
Containers inside pods partially share /procwith the host system include path about a memory and CPU information.
See also this article about Memory inside Linux containers.
I have Prometheus setup via Helm from Terraform and it's is configured to connect to my Kubernetes cluster. I open my Prometheus but I am not sure which metric to choose from the list to be able to view the CPU/MEM of running pods/jobs.
Here are all the pods running with the command (test1 is the kube namespace):
kubectl -n test1 get pods
podsrunning
When, I am on Prometheus, I see many metrics related to CPU, but not sure which one to choose:
prom1
I tried to choose one, but the namespace = prometheus and it uses prometheus-node-exporter and I don't see my cluster or my namespace test1 anywhere here.
prom2
Could you please help me? Thank you very much in advance.
UPDATE SCREENSHOT
UPDATE SCREENSHOT
I need to concentrate on this specific namespace, normally with the command:
kubectl get pods --all-namespaces | grep hermatwin
I see the first line with namespace = jobs I think this is namespace.
No result when set calendar to last Friday:
UPDATE SCREENSHOT April 20
I tried to select 2 days with starting date on last Saturday 17 April but I don't see any result:
ANd, if I remove (namespace="jobs") condition, I don't see any result either:
I tried to rerun the job (simulation jobs) again just now and tried to execute the prometheus query while the job was still running mode but I don't get any result :-( Here you can see my jobs where running.
I don't get any result:
When using simple filter, just container_cpu_usage_seconds_total, I can see the namespace="jobs"
node_cpu_seconds_total is a metric from node-exporter, the exporter that brings machine statistics and its metrics are prefixed with node_. You need metrics from cAdvisor, this one produces metrics related to containers and they are prefixed with container_:
container_cpu_usage_seconds_total
container_cpu_load_average_10s
container_memory_usage_bytes
container_memory_rss
Here are some basic queries for you to get started. Be ready that they may require tweaking (you may have different label names):
CPU Utilisation Per Pod
sum(irate(container_cpu_usage_seconds_total{container!="POD", container=~".+"}[2m])) by (pod)
RAM Usage Per Pod
sum(container_memory_usage_bytes{container!="POD", container=~".+"}) by (pod)
In/Out Traffic Rate Per Pod
Beware that pods with host network mode (not isolated) show traffic rate for the whole node. * 8 is to convert bytes to bits for convenience (MBit/s, GBit/s, etc).
# incoming
sum(irate(container_network_receive_bytes_total[2m])) by (pod) * 8
# outgoing
sum(irate(container_network_transmit_bytes_total[2m])) by (pod) * 8
I am trying to see how much memory and CPU is utilized by a kubernetes pod. I ran the following command for this:
kubectl top pod podname --namespace=default
I am getting the following error:
W0205 15:14:47.248366 2767 top_pod.go:190] Metrics not available for pod default/podname, age: 190h57m1.248339485s
error: Metrics not available for pod default/podname, age: 190h57m1.248339485s
What do I do about this error? Is there any other way to get CPU and memory usage of the pod?
I saw the sample output of this command which shows CPU as 250m. How is this to be interpreted?
Do we get the same output if we enter the pod and run the linux top command?
CHECK WITHOUT METRICS SERVER or ANY THIRD PARTY TOOL
If you want to check pods cpu/memory usage without installing any third party tool then you can get memory and cpu usage of pod from cgroup.
Go to pod's exec mode kubectl exec -it pod_name -n namespace -- /bin/bash
Run cat /sys/fs/cgroup/cpu/cpuacct.usage for cpu usage
Run cat /sys/fs/cgroup/memory/memory.usage_in_bytes for memory usage
Make Sure you have added the resources section (requests and limits) to deployment so that it can calculate the usage based on cgroup and container will respect the limits set on pod level
NOTE: This usage is in bytes. This can vary upon pod usage and these values changes frequently.
kubectl top pod <pod-name> -n <fed-name> --containers
FYI, this is on v1.16.2
Use k9s for a super easy way to check all your resources' cpu and memory usage.
As described in the docs, you should install metrics-server
250m means 250 milliCPU, The CPU resource is measured in CPU units, in Kubernetes, is equivalent to:
1 AWS vCPU
1 GCP Core
1 Azure vCore
1 Hyperthread on a bare-metal Intel processor with Hyperthreading
Fractional values are allowed. A Container that requests 0.5 CPU is
guaranteed half as much CPU as a Container that requests 1 CPU. You
can use the suffix m to mean milli. For example 100m CPU, 100
milliCPU, and 0.1 CPU are all the same. Precision finer than 1m is not
allowed.
CPU is always requested as an absolute quantity, never as a relative
quantity; 0.1 is the same amount of CPU on a single-core, dual-core,
or 48-core machine.
No, kubectl top pod podname shows metrics for a given pod, Linux top and free runs inside a Container and report metrics based on Linux system reporting based on the information stored in the virtual filesystem /proc/, they are not aware of the cgroup where it runs.
There are more details on these links:
Why top and free inside containers don't show the correct container memory
Kubernetes top vs Linux top
A quick way to check CPU/Memory is by using the following kubectl command. I found it very useful.
kubectl describe PodMetrics <pod_name>
replace <pod_name> with the pod name you get by using
kubectl get pod
You need to run metric server to make below commands working with correct data:
kubectl get hpa
kubectl top node
kubectl top pods
Without metric server:
Go into the pod by running below command:
kubectl exec -it pods/{pod_name} sh
cat /sys/fs/cgroup/memory/memory.usage_in_bytes
You will get memory usage of pod in bytes.
Not sure why it's not here
To see all pods with time alive - kubectl get pods --all-namespaces
To see memory and CPU - kubectl top pods --all-namespaces
As heapster is deprecated and will not be releasing any future releases, you should go with installing metrics-server
You can install metrics-server in following way:
Clone the metrics-server github repo: git clone https://github.com/kubernetes-incubator/metrics-server.git
Edit the deploy/1.8+/metrics-server-deployment.yaml file and add following section just after command section:
- command:
- /metrics-server
- --metric-resolution=30s
- --kubelet-insecure-tls
- --kubelet-preferred-address-types=InternalIP
Run the following command: kubectl apply -f deploy/1.8+
It will install all the requirements you need for metrics server.
For more info, please have a look at my following answer:
How to Enable KubeAPI server for HPA Autoscaling Metrics
If you use Prometheus operator or VictoriaMetrics operator for Kubernetes monitoring, then the following PromQL queries can be used for determining per-container, per-pod and per-node resource usage:
Per-container memory usage in bytes:
sum(container_memory_usage_bytes{container!~"POD|"}) by (namespace,pod,container)
Per-container CPU usage in CPU cores:
sum(rate(container_cpu_usage_seconds_total{container!~"POD|"}[5m])) by (namespace,pod,container)
Per-pod memory usage in bytes:
sum(container_memory_usage_bytes{container!=""}) by (namespace,pod)
Per-pod CPU usage in CPU cores:
sum(rate(container_cpu_usage_seconds_total{container!=""}[5m])) by (namespace,pod)
Per-node memory usage in bytes:
sum(container_memory_usage_bytes{container!=""}) by (node)
Per-node CPU usage in CPU cores:
sum(rate(container_cpu_usage_seconds_total{container!=""}[5m])) by (node)
Per-node memory usage percentage:
100 * (
sum(container_memory_usage_bytes{container!=""}) by (node)
/ on(node)
kube_node_status_capacity{resource="memory"}
)
Per-node CPU usage percentage:
100 * (
sum(rate(container_cpu_usage_seconds_total{container!=""}[5m])) by (node)
/ on(node)
kube_node_status_capacity{resource="cpu"}
)
An alternative approach without having to install the metrics server.
It requires you to currently install crictl into Worker Nodes where pods are installed. There is Kubernetes task defined in official doc.
Once, you have installed it properly you can use the below commands. (I had to use sudo in my case, but, probably may not be required depending on your Kubernetes Cluster install)
Find your container id of the pod sudo crictl ps
use stats to get CPU and RAM sudo crictl stats <CONTAINERID>
Sample output for reference:
CONTAINER CPU % MEM DISK INODES
873f04b6cef94 0.50 54.16MB 28.67kB 8
To check the usage of individual pods in Kubernetes type the following commands in terminal
$ docker ps | grep <pod_name>
This will give your list of running containers in Kubernetes
To check CPU and memory utilization using
$ docker stats <container_id>
CONTAINER_ID NAME CPU% MEM USAGE/LIMIT MEM% NET_I/O BLOCK_I/O PIDS
you need to deploy heapster or metric server to see the cpu and memory usage of the pods
You can use API as defined here:
For example:
kubectl -n default get --raw /apis/metrics.k8s.io/v1beta1/namespaces/default/pods/nginx-7fb5bc5df-b6pzh | jq
{
"kind": "PodMetrics",
"apiVersion": "metrics.k8s.io/v1beta1",
"metadata": {
"name": "nginx-7fb5bc5df-b6pzh",
"namespace": "default",
"selfLink": "/apis/metrics.k8s.io/v1beta1/namespaces/default/pods/nginx-7fb5bc5df-b6pzh",
"creationTimestamp": "2021-06-14T07:54:31Z"
},
"timestamp": "2021-06-14T07:53:54Z",
"window": "30s",
"containers": [
{
"name": "nginx",
"usage": {
"cpu": "33239n",
"memory": "13148Ki"
}
},
{
"name": "git-repo-syncer",
"usage": {
"cpu": "0",
"memory": "6204Ki"
}
}
]
}
Where nginx-7fb5bc5df-b6pzh is pod's name.
Pay attention CPU is measured in nanoCPUs where 1x10E9 nanoCPUs = 1 CPU
I know this is an old thread, but I just found it trying to do something similar. In the end, I found I can just use the Visual Studio Code Kubernetes plugin. This is what I did:
Select the cluster and open the Workloads/Pods section, find the pod you want to monitor (you can reach the pod through any other grouping in the Workloads section)
Right-click on the pod and select "Terminal"
Now you can either cat the files described above or use the "top" command to monitor CPU and memory in real-time.
Hope it helps
In case you are using minikube, you can enable the metrics-server addon; this will show the information in the dashboard.
If you exec into your pod, using sh or bash, you can run the top command which will give you some stats about resource utilisation that updates every few moments.
Metrics are available only if metric server is enabled or third party solutions like prometheus is configured. Otherwise you need to look at /sys/fs/cgroup/cpu/cpuacct.usage for cpu usage, which is the total cpu time occupied by this cgroup/container and /sys/fs/cgroup/memory/memory.usage_in_bytes for memory usage, which is total memory consumed by all the processes in the cgroup/container.
Also don't forget another beast called QOS, which can have values like Bursted, Guaranteed. If your pod appears Bursted, then it will be OOMKilled, even if it has not breached the CPU or Memory threshold.
Kubernetes is FUN!!!
I try to get CPU/Memory usage of the k8s Cluster Nodes via metrics-server API, but I found the returned values of metrics-server is lower than actual used CPU/Memory.
The output of kubectl top command : kubectl top nodes
The following is the output of the free command, from which you could see the memory usage is great than 90%.
Why the difference is so high?
kubectl top nodes is reflecting the actual usage of your Kubernetes Nodes.
For example:
Your node has 60GB memory and you actually use 30GB so it will be 50% of usage.
But you can request for example:
100 MB and have a limit 200MB memory.
This doesn't mean you only consume 0.16% (100 / 60000) memory, but the amount of your configuration.
I know this is an old topic, but I think the problem is still remaining.
To answer simply, the kubectl top command shows ONLY the actual resource usage, and it is not related to the request/limits configurations in your manifests.
for example:
you could obtain a 400m:1Gi (cpu/memory) usage for a specifique node while total requests/limits are 1.5:4Gi (cpu/memory).
You will observe enougth available resources to schedule but actually it will not work.
requests/limits are impacting directly node resources (resources reservation) but it does not means they are completly used (what kubectl top nodes is showing).
I was looking for ways to get details on disk utilization (mainly writes and Delete's) on a per POD level. I did get google advice such as cAdvisor/heapster etc but none of them talk about disk usage profiling from POD perspective.
Any help on this is greatly appreciated.
TIA!
Assuming the pods are running a linux variant you can do:
kubectl exec -it <pod> cat /proc/1/io
Returns info on the main process' IO defined here
You could then write a script to run the above command (or use the kuberentes API) for each pod of interest.