K
Q

How to get number of pods (available or terminating) in kubernetes?

June 12, 2019

I need to find the number of pods currently NOT terminated (can be available, terminating, etc.) in order to prevent deployment from starting if there are still some pods not terminated.

UPDATE: If there are no pods available, I should get "0". Is it possible ?

-- tumm
kubernetes

Similar Questions

Override env values defined in container spec
Can we see transfer progress with kubectl cp?
Kubernetes ExternalName Services
What is the purpose of kubectl proxy?
Deploy WAR in Tomcat on Kubernetes

4 Answers

June 12, 2019

You can try:

kubectl get pods --field-selector=status.phase!=Succeeded,status.Phase!=Failed

If you look at the Pod Phases you can see that this covers all possible pods where all containers are terminated (either failed or succeeded)

If you specifically want the count you could use a bit of 

jq
and use:

kubectl get pods --field-selector=status.phase!=Succeeded,status.phase!=Failed --output json | jq -j '.items | length'

This returns the # of pods that are not 

Terminated
.

-- Blokje5
Source: StackOverflow
November 26, 2022

If you are on Linux, you can use the command line tool 

wc
(wordcount) to count the lines of 
kubectl
(combined with the no headers option) like this:

kubectl get pods --no-headers | wc -l
-- Marridi
Source: StackOverflow
March 10, 2022

As an alternative to the 

jq
answer, here is a 
wc -l
answer

k3s kubectl get pods --output name | wc -l

This is somewhat more convenient in that 

jq
may not be always installed.

-- user7610
Source: StackOverflow
June 12, 2019

If your use case is making sure you only have a single instance/replica of your application running at any given time, maybe it's better to rely on built-in Kubernetes options to handle that.

You can configure your Deployment object using 

replicas
, 
maxUnavailable
and 
maxSurge
to control how many instances are created of your application.

-- Jose Armesto
Source: StackOverflow