Overview

Connect to Prometheus to:

• Extract custom metrics from Prometheus endpoints
• See Prometheus Alertmanager alerts in your Datadog event stream

Note: Datadog recommends using the OpenMetrics check since it is more efficient and fully supports Prometheus text format. Use the Prometheus check only when the metrics endpoint does not support a text format.

See the Prometheus metrics collection Getting Started to learn how to configure a Prometheus Check.

Setup

Follow the instructions below to install and configure this check for an Agent running on a host. For containerized environments, see the Autodiscovery Integration Templates for guidance on applying these instructions.

Installation

The Prometheus check is packaged with the Datadog Agent starting version 6.1.0.

Configuration

Edit the prometheus.d/conf.yaml file to retrieve metrics from applications that expose OpenMetrics / Prometheus end points.

Each instance is at least composed of:

When listing metrics, it’s possible to use the wildcard * like this - <METRIC_NAME>* to retrieve all matching metrics. Note: use wildcards with caution as it can potentially send a lot of custom metrics.

More advanced settings (ssl, labels joining, custom tags,…) are documented in the sample prometheus.d/conf.yaml

Due to the nature of this integration, it’s possible to submit a high number of custom metrics to Datadog. Users can control the maximum number of metrics sent for configuration errors or input changes. The check has a default limit of 2000 metrics. If needed, this limit can be increased by setting the option max_returned_metrics in the prometheus.d/conf.yaml file.

If send_monotonic_counter: True, the Agent sends the deltas of the values in question, and the in-app type is set to count (this is the default behavior). If send_monotonic_counter: False, the Agent sends the raw, monotonically increasing value, and the in-app type is set to gauge.

Validation

Run the Agent’s status subcommand and look for prometheus under the Checks section.

Data Collected

Metrics

All metrics collected by the prometheus check are forwarded to Datadog as custom metrics.

Note: Bucket data for a given <HISTOGRAM_METRIC_NAME> Prometheus histogram metric are stored in the <HISTOGRAM_METRIC_NAME>.count metric within Datadog with the tags upper_bound including the name of the buckets. To access the +Inf bucket, use upper_bound:none.

Events

Prometheus Alertmanager alerts are automatically sent to your Datadog event stream following the webhook configuration.

Service Checks

The Prometheus check does not include any service checks.

Prometheus Alertmanager

Send Prometheus Alertmanager alerts in the event stream. Natively, Alertmanager sends all alerts simultaneously to the configured webhook. To see alerts in Datadog, you must configure your instance of Alertmanager to send alerts one at a time. You can add a group-by parameter under route to have alerts grouped by the actual name of the alert rule.

Setup

1. Edit the Alertmanager configuration file, alertmanager.yml, to include the following:

receivers:
- name: datadog
  webhook_configs: 
  - send_resolved: true
    url: https://app.datadoghq.com/intake/webhook/prometheus?api_key=<DATADOG_API_KEY>
route:
  group_by: ['alertname']
  group_wait: 10s
  group_interval: 5m
  receiver: datadog
  repeat_interval: 3h

Note: This endpoint accepts only one event in the payload at a time.

2. Restart the Prometheus and Alertmanager services.

sudo systemctl restart prometheus.service alertmanager.service

Troubleshooting

Need help? Contact Datadog support.

Further Reading

• Introducing Prometheus support for Datadog Agent 6
• Configuring a Prometheus Check
• Writing a custom Prometheus Check