Weekly Airbyte pipeline health report in Notion

Every Monday at 8am, walk through the state of our Airbyte pipelines and post a written health report to a Notion database so the data team knows which connections need attention before the week gets going.

Start by calling Airbyte List Connections to enumerate every active connection across our workspaces. For each connection, call Airbyte List Jobs filtered to the last seven days to get its recent sync history. Where you need more context on a specific connection (source, destination, schedule, stream configuration), call Airbyte Get Connection Details.

Then analyse the data for the trends the raw Airbyte dashboard does not surface. Specifically look for:

• Connections whose failure rate is rising this week versus the prior seven days.

• Connections whose extracted-records volume dropped or spiked by more than 30% compared to their recent baseline.

• Connections whose average sync duration crept up week-over-week.

• Connections that did not run at all in the last seven days despite being active.

Draft the report with three sections. First, an executive summary: one short paragraph on overall pipeline health plus the two or three items that matter most. Second, an at-risk connections table: one row per flagged connection, columns for the connection name, the pattern you spotted (rising failures, volume shift, slower syncs, missed runs), and your best guess at the likely cause based on the connection details and job history. Third, a healthy but worth watching section: connections that are still green but showing early warning signs worth a glance.

Every active connection should appear somewhere in the report, even if only briefly noted as healthy in the executive summary. Do not silently drop any.

Finally, use Notion Create a Page to add a new page in the specified Notion database. Title it something like "Airbyte pipeline health — week of <Monday's date>" and put the report content in the page body, using headings for each section and a proper table for the at-risk connections.

The database id and any secondary Notion property values (owner, tags, status) should be configurable inputs on the workflow. The reader is a data team lead who wants proactive pipeline observability rather than a raw dashboard, so bias the writing toward likely causes and next actions, not just numbers.