Automation
Retry
Runs the step before it again when that step fails — a flaky API batch, a database that drops the odd connection — waiting between attempts, then passes its rows on.
When to use it
Use it right after a step that talks to another system and sometimes fails for reasons that go away on their own.
In a pipeline
File InputREST APIRetryExport CSV
Settings
| Setting | Type | Required | Default or hint |
|---|---|---|---|
| Attempts | number | — | 3 |
| Wait between attempts (milliseconds) | number | — | 5000 |
| Retry on | select | — | any |
Example
Three tries at a flaky API
- Attempts=3
- Wait between attempts (milliseconds)=5000
- Retry on=Network and timeout errors only
If the API answers on the second attempt, the run carries on as if nothing happened and the step's note says so. After a third failure the run stops as usual.
Worth knowing
- Attempts counts the first try: 3 means the step runs at most three times.
- Network and timeout errors only skips retrying mistakes that would fail again anyway (a wrong column name, a bad formula).
- Previews do not wait between attempts. Inside a Loop, each item gets its own attempts.
- To carry on even after the last attempt fails, put an Error Handler after the Retry.