Health Dashboard

RudderStack’s Health dashboard provides an intuitive UI to monitor all your Event Stream and Reverse ETL pipelines. It also provides realtime observability metrics for the tracking plans linked to your sources, including validation errors, violation types, etc.

To view this dashboard, log in to your RudderStack dashboard and go to Monitor > Health option in the left navigation bar.

Overview

In the Overview section, you get a quick summary of the destinations with failures across your Event Stream (both cloud and warehouse destinations) and Reverse ETL pipelines. You also see the events with violations for all the tracking plans linked to your sources.

You can filter these metrics by period - one day, one week, or one month - depending on your requirement.

Event Stream destinations

This dashboard shows the event delivery and failure metrics only for cloud mode connections. It does not include data for the device mode connections.

See Connection modes for more details on cloud mode and device mode connections.

In this view, you get a list of all the Event Stream destinations in your workspace with the following details (along with the change percentage):

• Events delivered (sortable by count or rate of change)
• Failures
• Failure rate

RudderStack provides a toggle to filter your destinations by Cloud and Warehouse. Click the Failures tab to view only the destinations that have event failures. By default, all the destinations are shown in this view.

You can also filter metrics for only the enabled/disabled destinations by clicking the filter option:

Destination-level event failure metrics

To get the destination-level data for event failures, click the row. A panel pops up on the right with details on the failed events.

Cloud destinations

RudderStack provides the following details for the failed events associated with the cloud destination:

• Event name
• Event type (identify, track, page, etc.)
• Source
• Count: Number of failed events.
• Last happened: When the error last occurred.

Click the event to see a sample failed event payload along with corresponding error details. For more details, click the View Destination button on the top right.

Warehouse destinations

RudderStack provides the following details for the failed events associated with the warehouse destination:

• Staging events: These correspond to the errors that occur in the staging process (during transformation or object storage, for example) before the syncs actually start. You will see the following details:

  • Event name
  • Event type (identify, track, page, etc.)
  • Source
  • Count: Number of failed events with the Event name.
  • Last happened: When the error last occurred.

Click the event to see a sample failed event payload along with corresponding error details. For more details, click the View Destination button on the top right.

• Syncs: These correspond to the errors that occur during the warehouse syncs. You will see the following details:

  • Error category
  • Source
  • Events count
  • Syncs count
  • First happened
  • Last happened
  • Status

Click the event to see a sample error. You can also retry syncing the event to the warehouse by clicking the Retry all button.

Change percentage calculation

RudderStack calculates the change percentage for the Event Stream destinations as follows:

Here, period is the time period by which you want to filter the metrics - one day, one week, or one month.

Reverse ETL connections

In this view, you get the following information on the latest syncs that are ongoing or completed across each Reverse ETL connection.

• Source-destination connection
• Status of the latest run (In progress, Completed without failures, Completed with failures, or Aborted)
• Duration of the sync
• Sync start time
• Failures (Percentage of deltas (new rows) that failed to sync)
• Invalids (Invalid records sent from the source)
• Summary of failed or aborted syncs in the selected duration (1 day, 1 week, or 1 month)

Each row corresponds to an individual connection with details on the latest sync and a summary of the failed or aborted syncs during the selected time period.

The Aborted status code implies an unsuccessful sync due to a number of reasons:

• Sync was aborted or stopped manually.
• RudderStack encountered issues while connecting to the warehouse due to incorrect configuration, changed credentials, or downtime.

Hover over the Failures column to see percentage of failed deltas (new records since last sync). In the below image, the latest run status is Completed, with failures as the deltas failed to sync.

Hover over the Invalids column to see percentage of invalid records sent from the source. In the below image, the latest run status is Completed, no failures as RudderStack did not face any errors or failures while syncing the deltas.

Get sync details

Click a row to get the additional sync-specific details like:

• Sync type (full/incremental)
• Number of rows in source
• Number of deltas (new data since last sync)
• Invalid records

Tracking plans

In this view, you get a list of all the sources connected to a tracking plan in your workspace with the following details (along with the change percentage):

• Tracking Plan
• Events validated (sortable by count or rate of change)
• Violations (sortable by count or rate of change)

RudderStack also provides a Violations tab to view only the sources that have tracking plan violations.

Validation error details

To get the validation error details, click the source. A panel pops up on the right with the following details:

• Event name
• Event type (identify, track, page, etc.)
• Events validated
• Events dropped
• Last occurred: When the error last occurred.

Use the Version dropdown to view the above metrics by tracking plan version. This is helpful if your tracking plan has undergone revisions recently.

Click the event to see the violation type. You can also click the violation type to see a sample payload and violation description.

For more information on the above metrics and violation types, see the Tracking plans guide.

Questions? Contact us by email or on Slack