Alerts in your own Grafana

Grafana's alerting engine queries xScaler like any other Prometheus or Loki data source, so rules you already have keep working. This page covers that route.

Start with Insights

Insights → Alerting already has alert rules, contact points, notification policies, silences, mute timings, inhibition rules and templates, with nothing to install. Use your own Grafana when you have alerting there already and want it in one place.

Set up Grafana alerts against xScaler

1. Add xScaler as a Prometheus data source

In Grafana, go to Connections → Data sources → Add data source → Prometheus.

Field	Value
URL	https://euw1-01.m.xscalerlabs.com
Authorization header	Bearer <token>
Custom header X-Scope-OrgID	<tenant-id>

HTTP headers config:

Authorization: Bearer <token>
X-Scope-OrgID: <tenant-id>

Save & test. You should see "Data source is working".

---

2. Create an alert rule

1. Go to Alerting → Alert rules → New alert rule
2. Set Data source to your xScaler data source
3. Write a PromQL expression in the query editor:

sum(rate(http_requests_total{status=~"5.."}[5m]))
/ sum(rate(http_requests_total[5m])) > 0.05

4. Set the threshold and evaluation interval (e.g. every 1m, for 5m)
5. Add labels (e.g. severity: critical) and annotations (summary, description)
6. Assign the rule to a folder and evaluation group

---

3. Configure notifications

1. Go to Alerting → Contact points and add a contact point (Slack, PagerDuty, email, etc.)
2. Go to Alerting → Notification policies and route alerts by label to the right contact point

---

Where the rules run

Rules created in your own Grafana are evaluated by that Grafana. It has to be running for them to fire, and it needs network access to the xScaler query endpoints.

xScaler's own ingest endpoints do not host a Ruler, so a rules file posted to https://euw1-01.m.xscalerlabs.com/prometheus/config/v1/rules answers Ruler is not enabled on this deployment. Rules live either in Insights or in your own Grafana or Prometheus.

Recording rules

Recording rules run in your own Prometheus or Grafana. Write the results back to xScaler with remote_write and query them like any other metric.

Common alert expressions

# High HTTP error rate (> 5%)
sum(rate(http_requests_total{status=~"5.."}[5m]))
/ sum(rate(http_requests_total[5m])) > 0.05

# High CPU usage (> 90%)
100 - (avg by (instance) (rate(node_cpu_seconds_total{mode="idle"}[5m])) * 100) > 90

# Low disk space (< 10% free)
node_filesystem_avail_bytes{mountpoint="/"}
/ node_filesystem_size_bytes{mountpoint="/"} * 100 < 10

# Pod not ready
kube_pod_status_ready{condition="true"} == 0

# Redis memory near limit (> 90%)
redis_memory_used_bytes / redis_memory_max_bytes * 100 > 90

# PostgreSQL too many connections (> 90% of max)
sum(pg_stat_database_numbackends)
/ scalar(pg_settings_max_connections) * 100 > 90

---

Moving rules into Insights

Rules move over one at a time. The query is the part that carries across unchanged, and the rest is a form:

In Grafana	In Insights
The query	The same PromQL or LogQL
The threshold expression	The condition row: reducer, comparison, threshold
for	Pending period
Labels and annotations	The same fields
Contact points and policies	Rebuilt once, then shared by every rule

See Coming from Grafana Alerting for the full mapping, and Alert rules for the editor.

Grafana Alerting docs

For Grafana's own alerting features, see the Grafana Alerting documentation.