Keep data calls stable even when a service fails off and on. The flow adds smart retries with a short wait and stops cleanly when it hits a set limit. It helps teams that rely on APIs with timeouts, rate limits, or random errors.

The run starts by setting a tries counter, then calls your action step that you will replace with your real API call. On success it moves to a success path. On error it checks for known errors by status code or message. If the error is not one you want to stop on, it waits for a set time, increases the counter, and tries again. A check decides if tries remain. If yes, it loops back to call the action again. If not, it raises a clear error that says the retry limit is reached.

Setup is simple. Swap the placeholder step with your API node and turn on the error output. Set max tries and the wait time. Define the filter so known errors can go to a separate path. Expect fewer manual restarts, faster recovery from outages, and cleaner logs. Use it for handling rate limits, short outages, or slow endpoints without babysitting every run.

• Manual start for safe testing before going live
• Tries counter that initializes and updates on each cycle
• Pluggable action step you replace with your API call and enable the error output
• Condition check to flag known errors by status code or message
• Wait step that pauses for a set time before the next attempt
• Loop control that checks remaining tries and routes to continue or stop
• Stop with a clear message when the retry limit is reached
• Separate paths for success and known error for clean follow up actions

• Reduce manual restarts from hours to minutes by letting retries run on their own
• Lower failed runs by pausing and retrying after short delays
• Prevent infinite loops with a firm retry limit and clear stop
• Handle rate limits and timeouts without extra hands on the system
• Standardize error handling across different API calls

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. Open the node labeled Replace Me and swap it with the real node that calls your API or service. Double click the node, enable the error output in Node Settings, and connect success to the Success path and errors to the Catch known error step.
3. Set the tries counter: confirm the Set tries node reads the incoming tries value and defaults to 0. This keeps the counter stable across loops.
4. Configure the error filter: in the Catch known error step, add conditions for status code or error text you want to mark as known. Route known errors to the Known Error path and other errors to the Wait step.
5. Adjust retry timing: open the Wait step and set the pause you want, for example 5 seconds. This helps with rate limits and short outages.
6. Update loop control: in Update tries confirm the counter increases by 1 on each cycle. In If tries left set your max tries number, such as 3.
7. Customize stop behavior: edit the Retry limit reached node message if you want a specific alert or text for logs.
8. Run a test: click Execute Workflow, force an error in your API call, and watch the tries field increase, the wait pause, and the loop continue until it succeeds or stops at the limit.
9. Validate the success path: trigger a successful call and confirm it goes to the Success step and ends without retries.
10. Troubleshoot common issues: if it never stops, check the If tries left condition. If it never retries, review the error output wiring and the Wait duration.