Troubleshooting logging in to the CLI tool using a token in a CI/CD environment.

Hello. We are looking to use an automated system for deploying GraphQL schemas and studio, but we have some problems logging in to the CLI tool. Is it possible to use a token for logging in, or is it another method?

Hi User, you can create a deploy token for this purpose. To do so, go to your project on manage.sanity.io and then Settings > API > Tokens > Add new token - make sure to choose “Deploy token” under rights 🙂

Hi again , thank you for the reply. Where would we use this token when we need to log in?

Are you able to set it up as an environment variable in your pipeline? What are you using for automation?

We are trying to use Octopus. We added the token to the environmental variables but we can't figure out how to use the token to log us in to the CLI tool, so we can use commands like

sanity deploy

,

sanity graphql deploy

automatically.

Try running the commands as follows:
(Using the relevant env var for your token)

Thank you very much ! 😃

It does not seem to work, unfortunately. When we try to provide the SANITY_AUTH_TOKEN before the sanity command it doesn't recognize it as a command, and when we provide it after the command it gives an error saying we have to use

sanity login

.Is there an argument option to provide token, say
?

As an alternative, does the following approach produce any different results for you?

this error message is returned, I've tried a few different variations of the command you sent. I am currently just trying the commands in my personal terminal logged out to replicate the CI environment.

Could you try creating a Read+Write token and running the commands with that one locally to test?

By for example running ?If so, it returned the same error message as above.

And this one with the R+W token?

Could you try creating a Read+Write token and running the commands with that one locally to test?

When I ran
with the R+W token as the <token>.

That’s usually an issue with the path variable. I guess you’re running PowerShell?

Okey, do you suggest that I should configure the path variable to sanity? Do you have any documentation to set it up correctly?

would love to know if you guys have any idea how to setup this in our CD environment with Powershell.I've tried a few variations in Powershell, like using

$

in front of the

$SANITY_AUTH_TOKEN

. Would also be nice to see some more documentation on this topic in the future. 🙂Cheers
🤓

Absolutely, User, we should give some attention to documenting CI/CD flows in general. It’s too confusing right now and there a bunch of different options each with specific challenges, mostly around authentication though.
Unfortunately I’ve never used PowerShell myself, so I can only assist from the sidelines here. I’ve seen it being a path variable issue, at least when that error message you shared above occurs. However, it usually throws the error for any

sanity <command>

in that case, not specifically for the

SANITY_AUTH_TOKEN=

portion. I’d be happy to check with the team to see if anyone’s more familiar with PS.

That would be lovely , thanks.Note that I'm not seeing the

not recognized

error message when not using the

SANITY_AUTH_TOKEN

in front of the command.

I figured it out. In Powershell you have to define the environment variable like so: and notice the

;

to separate the command from setting the variable. I was inspired by this Stack Overflow answer: https://stackoverflow.com/a/1422082 .Thanks for all the help
🤓

This is awesome, ! I’ll note this in bold for when we write up those docs on CI/CD. Thanks for sharing the solution - great catch 🎉

This is awesome, ! I’ll note this in bold for when we write up those docs on CI/CD. Thanks for sharing the solution - great catch 🎉