I get so many errors when i try to enrich the company. There is nothing wrong with the inputs in the fields. When I rerun the errored cells again, they complete successfully. Is there any form of retry mechanism I can implement?
Yes, you can implement a retry mechanism for company enrichment errors through the Auto-update feature. When setting up your RocketReach integration, you can enable the Auto-update option which refreshes your data every 24 hours. For persistent errors, try these troubleshooting steps first: 1. Clear your browser cache and refresh the page 2. If you see empty rows or display issues, this might be a front-end performance issue that can be resolved by clearing cache Clay also offers a scheduling feature that can automatically re-run your enrichments on a set schedule, which provides an "always on" solution rather than manually refreshing sources and re-running enrichments. If you continue experiencing issues, you may want to request human assistance for more specific troubleshooting.
That doesn't help. I want run retries only on errored cells not on the whole column
I don't want to be refreshing my browser cache. I want to keep it all automated.
Hey, sorry you’re running into this. Can you share the table URL where you’re seeing the issue? Just to clarify, we don’t currently have a built-in way to automatically retry only the errored cells—it has to be done manually. If this is still happening, or if it’s happening often, we can take a closer look on our end. Let me know if you have more questions.
im order to avoid this, how many items in a batch is recommended for a webhook post? And what is the ideal delay between batches?
Also GPT errors out. I have to run the cells again. Is there a way to implement a retry mechanism in the future?
Hey, Thanks for flagging these errors—I’ll pass them along to our engineers to take a closer look since they shouldn’t be happening. For webhooks, we recommend keeping batches to 10–20 items with a 1–2 second delay between them. That helps reduce the chance of timeouts or overloading the endpoint. Regarding the GPT errors, those typically mean you’ve hit your usage limit or the model timed out. There’s no built-in retry mechanism yet, but we’re aware this is a pain point and are exploring better error handling options for future updates. You can also track errors in the "Errored Rows" to re-run them Let me know if you have more questions.
Hey, Just to share an update—our team confirmed there was an incident earlier today that caused those errors. It’s been resolved, so the “Find People” search should now be working without issues. Let me know if you have more questions.
Okay thanks a lot! 🙂
I will implement that strategy