Hi,
I am trying to simulate a multi-tenancy Kibana setup with multiple roles and role mappings, and running into issues with Kibana making strange HTTP calls to itself (just like in this page Unable to update UI setting Request failed with status code: 403) while attempting to use certain index patterns.
My Setup:
-
Running RPM installations of Elasticsearch and Kibana 7.6.1 on the same VM
-
2 Elasticsearch Roles which give Read permissions to indices matching
"prefix-service1-*"
and"prefix-service2-*"
respectively. -
2 indices named
prefix-service1-2020-06-01
andprefix-service2-2020-06-01
-
3 index patterns
"prefix-*"
,"prefix-service1-*"
, and"prefix-service2-*"
-
A user that can authenticate/log in to Kibana, with both roles assigned successfully.
The Issue:
when querying from the Discover
page, with one of the more specific index patterns ("prefix-service1-*"
for example) documents from the appropriate index show up just fine.
But when using the more "generic" index pattern ("prefix-*"
) that matches multiple indices:
-
only documents from
prefix-service1-2020-06-01
are displayed -
An error pops up on the screen saying:
Unable to update UI setting \n Request failed with status code: 403
-
Using the chrome developer console I can see this happens right when an API call is made to http://<kibana_host>:5601/s/<space_name>/api/kibana/settings with a payload of:
{"changes" : {"defaultIndex":"********-****-****-****-************"}}
(where "*"s are a mix of either lowercase chars and numbers)
I would greatly appreciate any advice or instruction on how to get the "general" index pattern to show documents from all matching indices, and resolve the error/bad API call