Adapted from the official dashboard located in the official nginx-controller repo: https://github.com/kubernetes/ingress-nginx/tree/master/deploy/grafana/dashboards.
Haven't dig up why yet, but in my setup with Grafana 6.3.6 (where at least for the performance dashboard 6.6.0 is recommended) and the nginx-controller 0.25.1, the range vector selectors in time series of [1m]
and [2m]
where causing no data being returned.
I also missed being able to see the data per host (as in the host name declared by each Ingress) or the namespace that defined the Ingress. So I added new parameters and filters for these.
The result is these 2 adapted grafana dashboards so they work with my setup. Each dashboard lists the changes I made over the official ones
Made the following changes:
- Added a parameter for the
exported_namespace
so its possible to filter on the namespace where each individual Ingress is defined - Updated the
Ingress Percentile Response Times and Transfer Rates
graph to use a range of[5m]
(5 minutes) instead of[2m]
Made these changes:
- Added the same parameters as in their main dashboard, which are passed as filters to every graph expression.
- Grouped by host rather than path (since in my case we usually define Ingress with
/
as path, but different hosts) - Updated all the expression to use a range of
[5m]
(5 minutes) instead of[1m]
.