Prometheus integration for Splunk On-Call 🔗
The Splunk On-Call and Prometheus integration provides you with the real-time monitoring data of an open source time-series database and the collaboration tools you need to easily address system issues or incidents. Prometheus』 s time-series database and monitoring specializes in providing your team with real-time, live updating system metrics to more quickly show when errors occur or requests fail.
The Splunk On-Call and Prometheus integration is easily configurable to help you aggregate time-series data and respond to incidents in one centralized location. Prometheus integrates with Splunk On-Call to help you identify, diagnose, and resolve incidents in real-time, as well as conduct more thorough post-incident reviews.
Requirements 🔗
This integration is compatible with the following versions of Splunk On-Call:
Starter
Growth
Enterprise
The integration supports AlertManager 0.8.0 and Prometheus 2.0.0-beta 2.
Activate Prometheus in Splunk On-Call 🔗
From Splunk On-Call, navigate to Integrations, 3rd Party Integrations. Prometheus and select Enable Integration. Copy the Service API Key to use in the following steps.
Configure Splunk On-Call in Prometheus 🔗
Download AlertManager from the Prometheus website and configure it. Use the following code in the YAML configuration file for the AlertManager. Make sure to replace the api_key with the previously-saved Service API Key and change the routing_key to the routing key you want to use.
route:
group_by: ['alertname', 'cluster', 'service']
group_wait: 30s
group_interval: 5m
repeat_interval: 3h
receiver: victorOps-receiver
receivers:
- name: victorOps-receiver
victorops_configs:
- api_key: 558e7ebc-XXXX-XXXX-XXXX-XXXXXXXXXXXX
routing_key: Sample_route
state_message: 'Alert: {{ .CommonLabels.alertname }}. Summary:{{ .CommonAnnotations.summary }}. RawData: {{ .CommonLabels }}'
To use custom fields or a proxy url, use the following snippet as a template:
route:
group_by: ['alertname', 'cluster', 'service']
group_wait: 30s
group_interval: 5m
repeat_interval: 3h
receiver: victorOps-receiver
receivers:
- name: victorOps-receiver
victorops_configs:
- api_key:
routing_key:
entity_display_name: '{{ .CommonAnnotations.summary }}'
message_type: '{{ .CommonLabels.severity }}'
state_message: 'Alert: {{ .CommonLabels.alertname }}. Summary:{{ .CommonAnnotations.summary }}. RawData: {{ .CommonLabels }}'
custom_fields:
: '{{ .CommonLabels.eai_nbr }}'
# We must set a proxy to be able to send alerts to external systems
http_config:
proxy_url: 'http://internet.proxy.com:3128'
Start Prometheus from the command line and configure it to talk to the Alertmanager. In this example, prometheus.yml is the configuration file for Prometheus and “http://localhost:9093” is the instance of the Alertmanager that Prometheus is pointing to.
./prometheus -config.file=prometheus.yml -alertmanager.url=http://localhost:9093
Next, start the Alertmanager from the command line using the Alertmanager configuration file from earlier. In this example, alertmanager.yml is the name of the configuration file.
./alertmanager -config.file=alertmanager.yml
Alerts from Prometheus appears in the Alertmanager as they are generated.
If you don』 t want to wait for an alert to be generated by Prometheus, send a test message to the Alertmanager. The following example uses curl with an instance of Alertmanager:
curl -H “Content-Type: application/json” -d '[{“labels”:{“alertname”:“TestAlert”}}]' localhost:9093/api/v1/alerts