Skip to main content

Getting started

This walkthrough builds your first Nomaflow job — a simple nightly database refresh — in about ten minutes. By the end you'll have a job that fires on a schedule, runs a step against a connector, records its run history and streams its logs live. Everything happens in the Nomaflow page; no TOML editing required.

The example uses a SQL step because every install has a connector to test it with. The same workflow applies to Python, HTTP and LDAP Sync steps — only the form fields change.

Before you start​

Make sure you have:

RequirementHow to check
The framework is running.Open the application — you can sign in.
You have a connector to use as a target.Open Settings → Connectors — at least one connector should be listed. The framework's default pool always counts.
Your role has the job:* permission (or superuser).If the Nomaflow link is visible in the menu, you're set.

If any of those are missing, the framework getting-started guide covers the install path and Authentication → Roles and permissions covers the role wiring.

Step 1 — Open the Jobs page​

Click Nomaflow in the top navigation. The page that opens is the Jobs catalogue — every declared job, with its last-run state, next scheduled fire and a row of actions.

Nomaflow · JobsThe full catalogue of scheduled and on-demand jobs.＋ New Job📅 Schedule view↻ Reloadnightly-reporting-refreshSUCCEEDEDetlDaily refresh of the reporting store from the operational DB.⏱ 0 2 * * * · next: tomorrow 02:00 · last: 12 h ago · 3 steps▶ Run nowenabled(empty placeholder — your first job lands here)

The top toolbar carries the three actions you'll use most: New Job (open the job builder), Schedule view (a calendar of upcoming fires) and Reload (refresh the catalogue).

Step 2 — Create a job​

Click ＋ New Job. The Job editor opens with three sections: Identity, Schedule and Steps.

FieldSet it to
Idmy-first-job (URL-safe — letters, digits, hyphens, underscores).
Description"My first Nomaflow job — refreshes a reporting view."
Tagstutorial (optional — used to group jobs).
EnabledLeave on.

Save the identity block before you add steps. The job is now in the catalogue but does nothing yet.

Step 3 — Add a SQL step​

Inside the editor, click ＋ Add step and pick SQL Query from the type menu.

FieldSet it to
Namerefresh-totals (used as a label in the run history).
Typesql_query.
ConnectorPick any writable connector — the framework's default is fine for a smoke test.
QueryA short statement against your target — for the framework's default pool, SELECT now() is enough to prove the wiring.

Save the step. The editor returns to the steps list with one step on it.

Pick a real query when you can

A SELECT now() works as a smoke test, but a job that runs a meaningful query (a REFRESH MATERIALIZED VIEW, a daily counter update, a small INSERT … SELECT) gives you a useful result row in the history.

Step 4 — Set a schedule​

Back in the editor's Schedule section, the schedule field expects a 5-field cron expression (with an optional 6th field for seconds).

GoalCron
Every minute (good for a quick test).* * * * *
Every day at 02:00.0 2 * * *
Every Monday at 09:30.30 9 * * 1
Every hour at minute :15.15 * * * *
Manual-only (no schedule).leave the field empty

For a first test, set * * * * * and pick a timezone (e.g. Europe/Paris). The schedule field shows a live preview of the next three fires so you can sanity-check the expression before saving.

Save the job. Within a minute, the first run lands in the history.

Step 5 — Watch the run​

Back on the Jobs page, your my-first-job card now shows:

  • A RUNNING badge (green pill) when the run is in flight.
  • A SUCCEEDED or FAILED badge when it settles.
  • The last run timestamp updating each minute.

Click the badge — it opens the Run detail page for that run, showing the step timeline, the inputs and outputs of each step and the live log tail.

Nomaflow · Run detail · my-first-job · run_a8c…STEPSrefresh-totals✓ 12 msStarted: 14:02:01Finished: 14:02:01Trigger: cronStatus: SUCCEEDEDLIVE LOG14:02:01.083 INFO run started · trigger=cron14:02:01.085 INFO step refresh-totals · type=sql_query14:02:01.088 INFO connector=default · 1 row affected14:02:01.090 INFO step refresh-totals · SUCCEEDED · 5 ms14:02:01.091 INFO run SUCCEEDED · 12 ms total

The log streams live over Socket.IO while the run is in flight — close the browser, the run keeps going; re-open the page, the log resumes where you left off.

Step 6 — Switch to manual-only​

A job firing every minute will fill the history quickly. Once you've confirmed the wiring:

  • Open the job editor.
  • Empty the Schedule field.
  • Save.

The job is still in the catalogue and still ▶-runnable from the Jobs page, but no longer fires on its own. This is also the right mode for on-demand jobs — rebuilds, reports, one-off scripts an operator triggers manually.

Alternatively, use the enable toggle on the job card to disable the schedule without losing it.

Step 7 — Try a manual run​

Click ▶ Run now on your job card. Three things can happen:

What you seeWhen
The job fires immediately.The job has no parameters and a single step — there's nothing the operator could usefully configure.
A Run with parameters modal opens.The job has shared params, a Python step with op_kwargs, or more than one step.
The button is disabled with a spinner.A previous run is still in flight.

The Run with parameters modal is a powerful piece of the UI — it lets the operator override per-fire what would otherwise need a job edit. The full walkthrough lives in Jobs → Catalogue.

What you've just built​

BehaviourWhere it came from
The job fires on a schedule.The cron expression in the Schedule section.
The framework picks it up without a restart.Save in the job editor calls POST /admin/reload automatically.
The run lands in the history.Every fire creates a run row, every step a step_run row.
You can re-run from the UI.▶ Run now dispatches through the same step engine as the cron fire.
Logs stream live.The runner publishes to a Socket.IO channel the Run detail page subscribes to.

You haven't written any TOML, opened a terminal or restarted a service. The whole loop — define, schedule, run, observe — lives in the Settings UI.

Where to go next​

You want to…Read
Build the mental model — what's a job, run, step, schedule?Concepts.
Learn the catalogue + run-with-params modal in depth.Jobs → Catalogue.
Walk through the full Job editor with every field explained.Jobs → Create a job.
Pick the right step type.Steps.
See three end-to-end patterns.Workflow recipes.
Set up Slack / email / webhook alerts on failure.Notifications.