January 10, 2023: v0.6.5 Python, v0.6.10 dbt package

🔥 What’s new?

  • Alerts suppression 💤🔕

    • Don’t want to get multiple alerts if the same test keeps failing?
    • You can now configure an alert_suppression_interval, this is a “snooze” period for alerts on the same issue.
    • Docs here
  • Alerts filtering using selectors 🔘

    • You can now configure edr to send only selected alerts, using tags, owners or models selectors.
    • This enables you to send only the alerts you actually care about.
    • Docs here
  • Generate report for a single invocation 1️⃣

    • Use selectors to generate a pre-filtered report with the results you want to see.
    • Try it: edr monitor report --select last_invocation
    • Docs here
  • Granular time buckets in anomaly detection tests

    • Now you can set time buckets that fit the updates cadence of your tables, to get better anomaly detection.
    • The new config enables to bucket by hour | day | week and set the interval size as well.
      • Example: 6 hours buckets for volume test
      properties.yml
      tests:
        - elementary.volume_anomalies:
            timestamp_column: updated_at
            time_bucket:
              period: hour
              count: 6
      
    • This was a very popular request here on Slack, special thanks to Rodrigo Loredo for your help with the implementation!

💫 More changes

  • Following the breaking change we experienced in dbt 1.3.2, we added dbt release candidates versions to our CI. This will enable us to prevent such incidents in the future.
  • We are evaluating a change in how we upload dbt artifacts to improve performance and unlock more use cases. For now you will see a warning message that will not impact your usage.