Overview
Cron workflows run on schedule; webhook workflows can be run from here. Job code stays version-controlled — this page starts, pauses and inspects, nothing more.
The last fifteen executions, newest first. Open one for its steps; the raw OpenFn log sits behind a disclosure so it never crowds the page.
The measles dashboard — maps, district charts, pivots — lives in DHIS2 analytics. This page shows what the pipeline last wrote, so you can confirm the write-back landed.
Case-level line list pulled from the DHIS2 tracker, cleaned with cleanepi and validated with linelist. Names are never extracted.
Live config from the DHIS2 DataStore — this is how the workflows behave: which model, its parameters, what is extracted and where results are written. Edits apply on the next run, never to a run in flight.
Who each province's alert reaches. The alarm resolves recipients from the province's DHIS2 user group at run time — whoever is in that group with an e-mail address set. The address below is used only as a fallback, when a group is empty or no member has an e-mail. The severity words further down are the ones that go in the subject line.
Thresholds read from
openfn/jobs/write_predictions.js, the job that both writes the risk score to
DHIS2 and composes the e-mail — so this scale cannot drift from what recipients see.
Live status of the systems the pipeline depends on, checked from the panel backend on each load.
Who edited config or ran a workflow, and when. Recorded by the panel in the DHIS2 DataStore.