Running multiple kubectl commands in a bash script

Question

I'm trying to create a dashboard where I can see the status of my pods within my cluster. So I have a bash script that goes something like:

SERVICE_ONE=$(kubectl get pods | grep service-one | grep -E -- 'CrashLoopBackOff|Terminating|Error|Fail')
SERVICE_TWO=$(kubectl get pods | grep service-two | grep -E -- 'CrashLoopBackOff|Terminating|Error|Fail')
SERVICE_THREE=$(kubectl get pods | grep service-three | grep -E -- 'CrashLoopBackOff|Terminating|Error|Fail')

if [ ${#SERVICE_ONE} = 0 ]
then
  SERVICE_ONE_STATUS="GOOD"
else
  SERVICE_ONE_STATUS="BAD"

I then post these results to a GUI. The only problem is that I have almost 20 of these kubectl commands, which altogether, causes the program to take over 2 minutes to run. I tried doing something like:

KUBE_OUTPUT=$(kubectl get pods)
SERVICE_ONE=$(grep service-one <<< ${KUBE_OUTPUT} | grep -E -- 'CrashLoopBackOff|Terminating|Error|Fail')
SERVICE_TWO=$(grep service-two <<< ${KUBE_OUTPUT} | grep -E -- 'CrashLoopBackOff|Terminating|Error|Fail')
SERVICE_THREE=$(grep service-three <<< ${KUBE_OUTPUT} | grep -E -- 'CrashLoopBackOff|Terminating|Error|Fail')

But this doesn't work since the kube output isn't saved in the same format as when I run it plainly on my terminal (there's spacing issues, no newlines, etc.). I was wondering if there was a better way of executing multiple kubectl commands at once?

Don't run multiple commands: run just one with all details for all pods in json format and use other tools like jq to parse that long clob. — zerkms
– zerkms, Commented Nov 19, 2019 at 23:25
Ok I figured spewing all the information into jsonpath might be the way to go, but it seemed like it wouldn't keep the same format. I'll probably have to change my logic for a lot of my commands, but it's probably the way to go — olivercollins-inhome
– olivercollins-inhome, Commented Nov 19, 2019 at 23:50

samk · Accepted Answer · 2022-01-25 15:25:36Z

1

since, running kubectl command every time to get the pod details will take the time.

i tried kubectl get pods with -o=jsonpath something like below

$  VAR=$(kubectl get pods -o=jsonpath="{range .items[*]}{.metadata.name}{'\t'}{.status.phase}{'\n'}{end}"$'\r' | grep -E -- 'CrashLoopBackOff|T
erminating|Error|Fail|Running')

This will filter the result with NAME and STATUS of the pod. the output would be like

$ IFS=' '
$ echo $VAR
my-deployment-55bc8b77dd-8plsw  Running
my-deployment-55bc8b77dd-crlbx  Running
my-deployment-55bc8b77dd-jqqx4  Running
my-deployment-55bc8b77dd-mnm5k  Running
my-deployment-55bc8b77dd-zndvs  Running

Now you can again filter based in the pod name

$ SERVICE=$(echo $VAR | grep my-deployment-55bc8b77dd-8plsw)

the output will be

$ echo $SERVICE
my-deployment-55bc8b77dd-8plsw  Running

now split the result and display the status as you like. If this is not what you are looking for. plz put a comment.

edited Jan 25, 2022 at 15:25

answered Nov 20, 2019 at 7:41

samk

4406 silver badges13 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

olivercollins-inhome Over a year ago

I was really just missing this step:

VAR=$(kubectl get pods -o=jsonpath="{range .items[*]}{.metadata.name}{'\t'}{.status.phase}{'\n'}{end}"$'\r' | grep -E -- 'CrashLoopBackOff|T erminating|Error|Fail|Running')

, I ended up using echo ${VAR} | tr ' ' '\n' | awk '/service-one/{getline; print}' | grep -E -- 'CrashLoopBackOff|Terminating|Error|Fail' to check the status of each pod matching my service

samk Over a year ago

glad it helped!!:)

Collectives™ on Stack Overflow

Running multiple kubectl commands in a bash script

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related