Skip to main content
Data pipelines can sometimes fail. Although this is inherent to working with data and external systems, Nekt provides logs and alerts to help you understand and fix issues quickly.

​
Run logs

Each run has a log that records what happened during execution. Use logs to:
  • See when the run started and finished.
  • Understand which step failed (source, query, notebook, history, or destination).
  • Read error messages and stack traces when something goes wrong.
From the Runs page, click into a run to open its details and view the full log. For transformations (queries, notebooks, histories), logs often contain error messages from your SQL or code that point directly to the problem.

​
Alerts (email notifications)

Nekt lets you receive email notifications when runs fail so you can act quickly:
  • Per-resource alerts: enable alerts for individual sources, queries, notebooks, histories, and destinations from their details pages.
  • Global alerts: owners and admins can configure workspace-wide alerts in Settings → Alerts so one or more email addresses receive notifications for any failed pipeline.
For full details on how alerts work and how to manage them (including bulk actions), see Alerts.

​
Common failure reasons

Here are some common reasons why a run might fail and how to deal with them.

​
An adjustment is required

If something is wrong with a connector (for sources and destinations), a query or notebook, or any credentials used to access external tools, you will usually need to fix or update that configuration. The easiest way to understand the error is by checking the run logs. Logs typically include a message explaining what happened and what needs to be changed. If there is nothing useful logged, contact our support team and we will investigate further.
We are updating all resources to provide meaningful and actionable error messages, but there are many scenarios and we might not have covered all of them yet. We’re sorry for any inconvenience.

​
Instability or temporary issues

Pipelines can also fail due to temporary issues (e.g. network instability, upstream API timeouts, or cloud service limits). Nekt lets you configure retry parameters for each source, query, notebook, history, or destination to handle these cases gracefully:
  1. Number of retries: how many times to retry before marking the run as failed.
  2. Retry delay: how long (in seconds) to wait between retry attempts.
  3. Max consecutive failures: how many consecutive failures are allowed before inactivating the pipeline.
    • Use case: this is ideal for preventing repeated failures during off-hours or weekends, saving resources and avoiding unnecessary alerts.
These parameters can be managed in the Settings tab of each source, query, notebook, history, and destination, giving you flexibility in how you balance reliability, noise, and cost.