We upgraded Graylog from 3.1 to 4.1. We wrote a client that interacts with the API to export and import dashboards. We adjusted the API and it works mostly. We encountered an issue while POSTing a new dashboard to a fresh Graylog. The API states that the search_id is optional. We expect that an object will be created if the search_id is not in the POST. However, it seems to be mandatory.
Description of steps you’ve taken to attempt to solve the issue
Since we create our dashboards on a dev Graylog and afterwards export them to our production Graylog, we can retrieve the search object and POST it to our production Graylog before we POST the dasboard itself.
Hello,
I noticed there was not a reply. Maybe I can help.
From what I understand from this post is that you have a client that will post dashboard/s from your Dev GL server to your Production GL server, is this correct?
After upgrading this no longer works? or I think you stated it kind of works? Could you explain in greater detail about this?
According to the API, a POST against “/view” has an optional field search_id but instead it is mandatory. We assumed that a new search object would be created upon leaving search_id empty.
Thanks for the added info.
One of the reason I want to help you out was that in our environment we use mainly GET API’s. I haven’t got around to actually using POST API’s yet.
I’ve been banging my head against the wall over here for hours. I tried multiple ways of not using search_id but like your self I received the same error.
