Here’s something worth thinking about when you’re building automations between n8n and ServiceM8.
Imagine you’ve built a workflow that creates a follow-up task whenever a particular event happens in ServiceM8.
You test it. It works. Brilliant!
But what happens if the same event reaches your workflow twice?
Depending on how you’ve built it, you might end up with two tasks, two customer messages, or even two jobs.
A successful execution doesn’t necessarily mean a successful process.
One of the most useful habits you can build is to ask: What happens if this workflow runs again with exactly the same information?
For workflows that create things in ServiceM8, consider adding a check before the create step:
- Identify something unique about the event or record, such as a source record ID.
- Check whether you’ve already processed that ID.
- If you have, stop. If you haven’t, carry on.
- Record the ID once the operation succeeds.
There’s one extra complication worth knowing about: what if ServiceM8 successfully creates the task, but your workflow fails before recording that it has done so?
A simple check won’t protect against every duplicate. For important automations, you’ll want to think about retries, partial failures and simultaneous executions too.
You might hear this called idempotency. A slightly terrifying word, a very useful principle: running something twice shouldn’t create an unwanted second result.
Have you ever had an automation run twice when you only expected it to run once?
I’d be interested to hear what caused it!
~ Hazel
