Project

General

Profile

Feature #43668

mgr/dashboard: extend Grafana dashboard updates to support multiple Grafana instances

Added by Ernesto Puerta 29 days ago. Updated 28 days ago.

Status:
New
Priority:
Normal
Category:
dashboard/monitoring
Target version:
% Done:

0%

Source:
other
Tags:
Backport:
nautilus
Reviewed:
Affected Versions:
Pull request ID:

Description

Extend current:

ceph dashboard grafana dashboards update # implictly using the grafana instance configured in GRAFANA_API_URL

to:

ceph dashboard grafana dashboards update https://192.168.0.1:3000,https://192.168.0.2:3000,https://192.167.0.3:3000 # comma-separated list of addresses

History

#1 Updated by Lenz Grimmer 29 days ago

Can you please elaborate on the rationale for this change? What is the goal of supporting multiple Grafana instances?

#2 Updated by Ernesto Puerta 28 days ago

Lenz Grimmer wrote:

Can you please elaborate on the rationale for this change? What is the goal of supporting multiple Grafana instances?

Sorry for the lack of context: this comes from discussion with OpenStack team. They are deploying collocated ceph-dashboard+Prometheus+Grafana and want to achieve HA on all these components. Therefore, Grafana is routed via HAProxy (this could also work with a ceph-mgr localized setting).

Nevertheless the idea was to keep the legacy behavior when no URL is defined.

Also available in: Atom PDF