SaaS REST API - Automations endpoints questions (documentation issues?)

[stefanstoichev123]

Looking at the Automations API and have few questions:

• what is the difference between automation, id and automationId in the path parameters?
• what is the difference between run and runId in the path parameters?
• (probably because of the first point) why are there duplicate endpoints? for example
  • /v1/automations/{automation}/actions/copy and /v1/automations/{id}/actions/copy
  • /v1/automations/{automation}/runs/{run}/actions/retry and /v1/automations/{automationId}/runs/{runId}/actions/retry
• some endpoints are returning id and some guid (as automation id). Which is the correct one? For example:

        

         and

       

Is all of the issues just documentation ones or there are difference endpoints (for some reason)?

Regards!

Stefan

[DaveChannon · Accepted solution]

Thanks for posting this @stefanstoichev123 

Appreciate that this was a while ago, so for others who might find this post please note that we have improved the documentation for this API and will continue to do so.

GUIDs now shows as IDs, with the exception of the usage endpoint.

I've raised another improvement to tidy up the {automationId} on runs into {id} to align with the rest of the endpoints on https://qlik.dev/apis/rest/automations.

Thanks again

[DaveChannon · Accepted solution]

All endpoints have now been updated and are consistent on https://qlik.dev/apis/rest/automations 

Thanks again for the feedback!

[stefanstoichev123]

Great! Thank you very much @DaveChannon 

Is there any official way to follow the SaaS API changes in details? I have an workaround for myself but just asking 🙂

Also ... here is the best way to report issues (if any) with SaaS API?

Stefan

[DaveChannon]

No out-of-the-box method as yet. As you've alluded you could watch the API specs in the meantime but we know we need to make it easier to follow 🙂 Do you have any examples of companies who make it really easy to follow API changes?

For reporting issues, best way is a support case through the support portal. We do try and monitor the community here too for those who don't have access to the portal, but there's no guarantee that it'll get picked up.

[stefanstoichev123]

Personally I went the "dummy" way - web scrape qlik.dev and download the specs for each endpoint. Then combine the specs and create a valid swagger schema and commit/push to a repo (all executed by GitHub Action).

Its not nice but its working for now 🙂

Stefan

[DaveChannon]

Functional is good! 😉