Timeout is to short, or needs to be configured to monitor for running processes.
I have been trying to generate a listing of all of the blob sizes in a container, and it takes longer than the timeout settings and will not complete. If I run against a small container it works properly. Also, while it states I have been idle for 20 minutes I don't believe it was actually that long.

50 comments
-
Mark S commented
Agreed. Also, the timeout is sometimes not extended when input is entered. I just edited a file and was creating some resources via Terraform and it "timed out" about 5 minutes after I started creation.
-
Anonymous commented
Please increase it to 1h
-
Anonymous commented
I need it for long running processes
-
Vincent Macchia commented
Any update on this? it's been 3 years!
-
COG commented
I agree with this. IT says idle for 20mins meanwhile it logs out sporadically. This one time, I took my eyes out for a minute and it logged out. My gripe is that when you log back in it takes you back to the home folder. please fix. I understand the timeouts in the portal, but the timeouts in the shell.azure.com is unacceptable.
-
Scott Shorkey commented
To add to the chorus, it's needed to create kubernetes clusters, as that typically takes longer than 20 minutes.
-
Vatan Joshi commented
I have to start my from scratch every time just because of this. For me its not even staying 20 min active. This is too sad. I know you guys can do better.
-
aidman_nz commented
I find cloudshell shuts down frequently due to 'inactivity' while I have a powershell script or terraform deployment running even when it has been running for <10mins and not for 20 mins like the timeout threshold is supposed to be. Please fix this asap as it needs to be robust for long running tasks and not just for running quick one off commands.
-
Alex Kolisnychenko commented
We're running "az vmss update-instances" command and it can take forever to finish, depending on the cluster size. 20 minutes is not adequate timeout for the shell.
-
Richard commented
this timeout is terrible. We are running a long script that creates Virtual Network Gateways so the timeout means we dont know how the script is progressing
-
Anonymous commented
Needful change, indeed very annoying when running comprehensive commands.
-
Trung Lai commented
Very annoying! Please increase the timeout or provide it as a configurable option.
-
Jamey Patterson commented
If you're in the middle of running a process that takes excess of the timeout period,
Is it possible to actually detect what is idle?
If not processing anything, actually idle, then timeout is fine.
But if you are waiting for a process to finish, please don't timeout. -
Ekant commented
Please increase timeout value to at least 1 hour.
-
Tab S commented
I also added this comment to another similar thread: Adding a configurable timeout length would be great. But at least if it is possible for Microsoft to extend the timeout, please extend it to at least an hour. I have created scripts that automate deployment of a lot of resources that altogether takes up to that long to complete (it includes creating and setting up an API Management Service, which by itself takes more than 20 minutes). And we have a need to use the Cloud Shell for this for 2 reasons: multiple people would have access to the script files and the platform, and success of the deployment is not dependent on a particular user's system. We need this for the sake of regulation and compliance.
-
Anonymous commented
I've been booted out after 10 minutes despite not actually being idle. 1/2 way through typing a command in and bang. see ya. start again please.
Very very very frustrating.
-
Oleh Romanchuk commented
Well, it's just frustrating, timed out during terraform deployment. It's just too buggy so far.
As a workaround, I'd suggest trying VSO https://online.visualstudio.com (worked for me)
Also, I'm going to try out tmux + some key-stroke generating script and see if it helps.
Regardless of the time out issue, I have a suggestion/request. It'll we very useful if you could make disown work in azure cli.
https://linux.die.net/man/1/disown -
Nipsy commented
So is there currently no way to know what size each of my containers is?
I have a few containers totalling around the 2TB mark, so it takes more than 20 minutes for the Azure Cloud Shell to calculate each of their sizes. I would like to know how big each one is but I guess it's just impossible...
-
Tom J Sweet commented
Times out during the https://www.azuredevopslabs.com/labs/vstsextend/kubernetes/ tutorial
-
Reiji Saito commented
"Unset autologout" does not work.
Provisioning including "Application-Gateway" may take more than one hour in some cases.
The terminal session is disconnected even though the task of "Ansible-playbook" is being executed. . .
Aren't Microsoft developers too stupid to users?