Datadog migration considerations

Datadog uses its own proprietary query language and storage system. Chronosphere is based on Prometheus. As you consider moving from Datadog to Chronosphere, review the following questions and suggestions as you plan your migration.

General

• Are you planning to stay on DogStatsD, or move to Prometheus (opens in a new tab) or Open Telemetry (opens in a new tab)?
• How's your cluster defined?

Metrics

Understanding the metrics you're using and how they're used can help determine the time needed to switch from the previous system to Chronosphere.

• Do you have a list of the metrics you are using for dashboards and alerts?

• How many dashboards do you need to convert?

  • How many widgets per dashboard?
  • Determine the breakdown of widget type (visualization) per dashboard.

• Which dashboards are critical to you and your users?

  You can use popularity ranking in Datadog (opens in a new tab) to determine this:

  • Get All Dashboards (opens in a new tab).
  • Iterate over these dashboards' IDs: Get a dashboard (opens in a new tab).

• How many monitors do you need to convert?

  • How do you route alerts in Datadog today?
    • Dynamic routing based off of labels.
    • Static routing per monitor.
  • Which monitors are critical? Monitors/alerts that are going to PagerDuty are often the most critical. Example: Get All Monitor Details (opens in a new tab).

Integrations

Chronosphere integrates with other systems.

• What integrations are you currently using?
  • What metrics do those integrations use, either ingested or exposed.
  • How do you use those metrics in your dashboards and alerts? This helps create the correct exporter configuration. For example, configuring YACE (opens in a new tab) to do a sum statistics aggregation versus average gives different results.
• Do you have a plan for how to ingest the OSS metrics?
• Get consensus about using OSS Grafana dashboards for OSS metrics, or if you'll need to convert the widgets.

Custom metrics

You and your users might have created custom metrics for internal applications.

• How many metrics are distributions? Egress costs might be higher for distribution metrics than for other types.
• Develop an understanding of what your custom metrics are, what they're monitoring, and what labels they have.

Alerts migration

Alerts and notifications are often critical parts of your observability platform.

• Obtain a list of the existing alerting platforms being used.
• Create a list of the notification channels in use.