Schedule automations and trigger them by API

What can start an automation

Scheduling controls the timing of runs, such as every day at 6:00 AM, weekly on Mondays at 9:00 AM or monthly on the 1st. A trigger is what starts a run.

Choose the right trigger

Where you set the schedule for each type

To change the schedule of a saved automation, open it for editing and click Edit schedule in the Schedule row. See Manage automations.

Set how often an automation repeats

1. Pick a unit: Day, Week, Month or Year.
2. Next to Repeat every, pick the number of units between runs.
3. For Week, also pick the weekdays to run on.

New schedules offer Day and coarser only, so you can't create an hourly or every-few-minutes schedule. If an automation was saved earlier with an Hour or Minute schedule, the scheduler still shows that unit when you edit it, so you can see what is running.

Run an automation weekly on chosen days

When Week is selected, a row of weekday buttons appears (one letter per day; hover for the full name). Select one or more days. At least one day is required. The first run can only fall on one of the days you pick.

Run an automation monthly or yearly

A monthly schedule runs on the day of the month of its first run. If that day doesn't exist in a month (for example the 31st), Mammoth shows a note such as Months without a 31st day are skipped. A yearly schedule that starts on 29 February shows Runs only in leap years, since other years have no Feb 29.

Plan limits on how often an automation runs

Your plan sets how often data can be refreshed. The scheduler doesn't offer a unit more frequent than your plan allows, whichever automation type you are creating. For example, on a plan whose limit is weekly, only Week, Month and Year are offered. On a plan without scheduled refresh, no unit is offered, so you can't set a schedule at all.

For Dataset Refresh and File Collection, Mammoth also checks the schedule:

• When you save, a schedule more frequent than your plan allows is refused with "Following frequency not allowed in your plan, please upgrade the plan".
• When a saved schedule is due but is now more frequent than your plan allows (for example after a plan change), that run is skipped.

Scheduled Export has no plan feature of its own, but the scheduler still offers only the units your plan's refresh frequency allows. Data Consolidation has no schedule, so it is never limited. The limits for each plan are on Limits and quotas.

Set the first run and an end date

1. Under First run, pick the date and time of the first run. Later runs repeat from it. You can't pick a past date. If the time you pick is less than 5 minutes away or already gone today, the first run moves to the next day at that time (for a Week schedule, the next weekday you selected).
2. To stop the schedule on a date, click + Add end date and pick an End date. It can't be earlier than the first run. To remove it, click Remove end date.

Without an end date, the automation runs until you suspend or delete it.

Below the dates, Next runs lists up to the next 3 run times, so you can check the schedule before you save. If the end date leaves only a few runs, the list shows All runs (or Only run when there is only one).

Messages that stop you continuing

You can't continue until the schedule is complete. The step tells you what is missing:

Which time zone automation schedules use

Dates and times you pick are in your browser's local time zone. There is no separate time zone setting. Mammoth stores the schedule in UTC and shows next and last run times in your local time.

Because the run time is fixed in UTC, its local time can move by an hour when daylight saving time starts or ends. A run close to midnight can also fall on a different local day than you expect. The Next runs preview already shows the local times the runs will use.

Example schedules

Stagger and sequence schedules

• Stagger concurrent automations. Rather than every refresh at 6:00 AM, use 5:30, 6:00 and 6:30 AM to avoid contention.
• If one process depends on another, schedule it at least 15 minutes after the first one usually finishes. For example: Dataset Refresh at 5:00 AM, pipeline at 5:15 AM, Messaging at 6:00 AM.
• Run non-urgent automations outside business hours so they don't slow down people working in Mammoth.
• Name automations with their timing, such as Weekly Executive Summary - Mon 8am.

To refresh several datasets in a fixed order, put them in one Dataset Refresh and choose In a sequence. See Refresh several datasets in parallel or in a sequence.

Run an automation now

Run any Active automation straight away: click Run now in the detail panel, the edit window or the three-dot menu. Mammoth confirms with "Automation triggered successfully". Run now is only available while the automation is Active. A suspended automation must be resumed first. See Run an automation now.

Use a manual run to:

• test a new automation before its schedule starts;
• re-run after fixing a failure;
• send data to stakeholders who need it now;
• check that a configuration change worked.

A Dataset Refresh set to Manual only has no automatic runs at all. It runs only when you trigger it manually or through the API.

Run an automation when a file arrives

Data Consolidation automations run when new files are uploaded to, or moved into, their source folder. No schedule or manual trigger is needed: the new file starts the consolidation. See Combine incoming data.

Receive data pushed by webhook

A webhook lets an external system push data into a Mammoth dataset over HTTP. Your application sends an HTTP POST with a JSON or form-data payload to the dataset's unique URL, with the secret in an X-SECRET header if the webhook has one. Mammoth holds the rows and adds them at the next Add new rows run (5 min, Hourly, Daily or By hand), or when you click Add now.

To create one, click New dataset in the Data Library, choose Connectors, then Push > Webhook. Sends with a wrong or missing X-SECRET are turned away and nothing from them is added. Keep the secret safe: anyone with the URL and secret can push data into the dataset. For every option on the Webhook card, see Receive data by webhook.

Scheduled Export as a trigger

A Scheduled Export runs the export steps you pick on its own schedule (Day, Week, Month or Year), whether or not the data changed. Those exports still run after their data updates too, unless you turn that off on the export.

Run an automation from another system with the API

External systems can run an existing automation through the Mammoth API, the same as clicking Run now. Use it when an upstream process should start Mammoth as soon as it finishes.

Run an automation with the REST API

Send a PATCH request to the automation with a command operation on the run path:

curl -sS -X PATCH \
  "https://app.mammoth.io/api/v2/workspaces/$WORKSPACE_ID/projects/$PROJECT_ID/automations/$AUTOMATION_ID" \
  -H "Authorization: Bearer $MAMMOTH_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"patch": [{"op": "command", "path": "run", "value": ""}]}'

To find the automation's id, list the project's automations with GET /workspaces/22/projects/5/automations.

The same endpoint suspends or resumes an automation. Send {"op": "replace", "path": "status", "value": "suspend"} (or "restore" to resume it) in the patch list.

A suspended automation can't be run through the API either. The request is refused with "Automation is suspended; resume it before running". Resume it first.

Run an automation with the CLI

With the mammoth CLI, the same run is:

mammoth automation update 123 --input '{"patch": [{"op": "command", "path": "run", "value": ""}]}' --yes

mammoth automation list shows the ids. See mammoth automation for every automation command.

Integration patterns

The API applies per-minute rate limits (see Limits and quotas). Retry rate-limited calls after a short delay instead of sending them again and again.

Troubleshooting

Related

• Create an automation: where the schedule step sits in the wizard.
• Manage automations: read Next run, Last run and statuses, and edit a schedule.
• Refresh a dataset: schedule refreshes for connector and From URL datasets.
• Combine incoming data: the file-arrival trigger.
• Receive data by webhook: every option on the Webhook card.