Bug 1600872 - For two prometheus instances only one user can be inserted.
Summary: For two prometheus instances only one user can be inserted.
Keywords:
Status: CLOSED WONTFIX
Alias: None
Product: OpenShift Container Platform
Classification: Red Hat
Component: Monitoring
Version: 3.9.0
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
: ---
Assignee: Paul Gier
QA Contact: Junqi Zhao
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2018-07-13 08:57 UTC by Priyanka Kanthale
Modified: 2021-09-09 15:01 UTC (History)
2 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2018-07-31 07:49:31 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Description Priyanka Kanthale 2018-07-13 08:57:28 UTC
Description of problem:

Two prometheus instances running (with separate service accounts) for failover reasons. 
Changed the 'prometheus-metrics-viewer' rolebinding to contain both service accounts. 
Added both the user in the 'passwd.yml' file in the logging-elasticsearch secret.
But this did not help for the second prometheus instance 

Actual results:
 Can only insert one user

Expected results:
 Both users can we inserted. 


Additional info:
Customer reports: 
 Encountered the last piece which is the 'sg_role_mapping.yml' file. Looks like the PROMETHEUS_USER env variable from the elasticsearch DC gets inserted into the role mapping file when the elasticsearch container starts [1].

So far he solved it by using the bearer token from the first prometheus instance in the second one.

[1] https://github.com/openshift/origin-aggregated-logging/blob/master/elasticsearch/run.sh#L82

Comment 1 Simon Pasquier 2018-07-31 07:49:31 UTC
I'm closing this as "Won't fix" because it impacts only the tech preview stack. By the way, running redundant Prometheus instances is usually achieved with a replica count > 1 rather than running 2 separate deployments.

Prometheus in OpenShift 3.11 will be deployed by the cluster monitoring operator which will run multiple Prometheus instances out-of-the-box for redundancy.


Note You need to log in before you can comment on or make changes to this bug.