Dashboards
Whole-Cluster Dashboard
Note: This dashboard is packaged by default with Gloo Edge Enterprise
A dashboard showing whole-cluster metrics can be found in the gloo/Envoy Statistics dashboard. In that dashboard you can find panels showing:
- Requests per Second
- Percent Response Code per Second
- Response Codes per Second
- Upstream Request Time Percentiles
- Round Trip Time Percentiles
- Average Request Time
Dynamically Generated Dashboards
Gloo Edge Enterprise’s observability component generates a dashboard per watched upstream. These dashboards are updated and recreated every time an upstream changes. It renders a Go template that can be found in a configmap, which gets loaded into the observability pod’s env at startup:
~ > kubectl -n gloo-system get cm gloo-observability-config -o yaml | head -n 10
apiVersion: v1
data:
GRAFANA_URL: http://glooe-grafana.gloo-system.svc.cluster.local:80
UPSTREAM_DASHBOARD_JSON_TEMPLATE: |2
{
"annotations": {
"list": [
{
"builtIn": 1,
If you want to customize how these per-upstream dashboards look, you can provide your own template to use by writing a Grafana dashboard JSON representation to that config map key. Currently the only available variables that are available when the Go template is rendered are:
.NameTemplate.Uid.EnvoyClusterName