The Monitoring page is accessible on CockroachDB dedicated clusters. This page allows you to:
- Set up cluster monitoring with Datadog.
- Access the cluster's built-in DB Console to view time-series data on SQL queries, troubleshoot query performance, view a list of jobs, and more.
To view the Monitoring page, log in and click Monitoring on the left-hand navigation.
Monitor CockroachDB dedicated with Datadog
To set up Datadog monitoring with CockroachDB dedicated, your Datadog account must be associated with a Datadog organization.
Enabling the Datadog integration on your CockroachDB dedicated cluster will apply additional charges to your Datadog bill. Your CockroachDB dedicated bill is unchanged.
For more information about using Datadog, see the Datadog documentation.
To enable Datadog monitoring for a CockroachDB dedicated cluster:
On the cluster's Monitoring page, click Setup in the Datadog panel.
Fill in the API key and Datadog Site fields with their corresponding values.
- The API key is associated with your Datadog organization. If you don't have an API key to use with your CockroachDB dedicated cluster, you need to create one. For instructions, see the Datadog documentation.
- Your Datadog Site corresponds to your Datadog Site URL. For more details, see the Datadog documentation.
Click Create. Depending on the size of your cluster and the current load on the system, the integration might take some time to become enabled.
Once it is registered on Datadog, the cluster will appear on your Datadog Infrastructure List. This can take up to several minutes.
Verify integration status
Once enabled, the Integration status in the Datadog panel on the Monitoring page will show as
If an issue is encountered during the integration, one of the following statuses may appear instead:
Activeindicates that the integration has been successfully deployed.
Inactiveindicates that the integration has not been successfully deployed. Setup has either not been attempted or has encountered an error.
Unhealthyindicates that the integration API key is invalid and needs to be updated.
Unknownindicates that an unknown error has occurred. If this status is displayed, contact our support team.
Metrics export from CockroachDB can be interrupted in the event of:
- A stale API key. In this case, the integration status will be
Unhealthy. To resolve the issue, update your integration with a new API key.
- Transient CockroachDB unavailability. In this case, the integration status will continue to be
Activebut you might experience incomplete metrics exports in Datadog. To resolve the issue, try deactivating and reactivating the integration from the Datadog panel. If this does not resolve the issue, contact our support team.
Gaps in metrics within Datadog do not necessarily point to an availability issue with CockroachDB. If you encounter any gaps in metrics, we recommend contacting support.
To monitor the health of metrics export, you can create a custom Monitor in Datadog.
View and configure dashboards
Open your Datadog Dashboard List and click
CockroachDB Dedicated Overview. This sample dashboard presents a high-level view of SQL performance and latency, and information about resource consumption to help aid in capacity planning. It provides the ability to drill down to specific nodes (identified by a
(node, region) tag pair) within your cluster.
Metric values and time-series graphs in Datadog are not guaranteed to match those in the DB Console, due to differences in how CockroachDB and Datadog calculate and display metrics.
Enable percentiles for selected metrics
A subset of CockroachDB metrics require that you explicitly enable percentiles for them in the Datadog interface. Graphs that display data for these metrics will fail to render properly otherwise.
To enable percentiles for these metrics, perform the following steps:
- In your Datadog interface, select Metrics then Summary from the sidebar.
- Check the Distributions checkbox in the Metric Type section to limit returned metrics to the subset of CockroachDB metrics that require percentiles support.
- For each metric shown:
- Select the metric and expand its Advanced section.
- Click the Edit button.
- Click the slider labeled Enable percentiles and threshold queries.
- Click the Save button.
You only need to perform this once per metric. Datadog graphs reliant on these metrics will begin rendering immediately once configured in this manner.
Only data received for these metrics once percentiles are enabled can be displayed; any data collected before enabling percentiles for these specific metrics cannot be rendered.
To preview the metrics being collected, you can:
- Click on your cluster's entry in the Infrastructure List to display time-series graphs for each available metric.
- Use the Metrics Explorer to search for and view
Monitor health of metrics export
To monitor the health of metrics export, we recommend that you create a new Monitor.
Select Threshold Alert as the detection method, which configures an alert that is sent when a supported metric reaches a given threshold. For descriptions of some useful CockroachDB alerts, see Monitoring and Alerting.
To Define the metric:
- Select the
- Export the metric from your CockroachDB dedicated cluster (the cluster name in the Infrastructure List.
- Select the
To Set alert conditions:
- Trigger when the metric is
on averageduring the last
- Set Alert threshold to
Notifyif data is missing for more than
- Trigger when the metric is
This monitor will notify your organization if Datadog is no longer receiving data from your CockroachDB dedicated cluster.
To update the metadata associated with the integration (for example, to rotate API keys):
In the Datadog panel, click the ellipsis and select Update.
Update the API key and Datadog Site fields and click Create. The integration will be redeployed.
To deactivate the integration:
In the Datadog panel, click the ellipsis and select Deactivate integration.
When disabled, the Integration status in the panel will show as
After deactivating an integration, the metrics data will remain in Datadog for a default retention period.
Access the DB Console
To access the DB Console:
On the cluster's Monitoring page, click Open DB Console in the DB Console panel.
You can also access the DB Console by navigating to
https://<cluster-name>crdb.io:8080/#/metrics/overview/cluster. Replace the
<cluster-name>placeholder with the name of your cluster.
Log in with your SQL username and password.
For details on creating additional users that can connect to the cluster and access the DB Console, see User Management.
Explore the DB Console
- For an overview of all the areas of the DB Console, see DB Console Overview.
- Be sure to check out the Node Map, which visualizes the geographic configuration of your cluster on a world map and provides real-time cluster metrics, with the ability to drill down to individual nodes. This Enterprise feature has been pre-configured and enabled for you.
If you have a single-node cluster, you may see a warning that you have under-replicated ranges. This is expected because the default replication factor is set to 3, and you can only have one replica per node. For more information about replication issues, see Cluster Setup Troubleshooting.