This morning I sent 7500 records from HubSpot to BIKA. I did 500 records yesterday with no issues at all.
Today after the 1st 150 or so I started to get lots of automation failure emails (about 1200 of them) with the following message:
Automation「From xxxx」 run failed
error message: Invalid prisma.database.findUnique() invocation: Timed out fetching a new connection from the connection pool. More info: Connection pool | Prisma Documentation (Current connection pool timeout: 10, connection limit: 9)
Does anyone have any ideas what went wrong?
There is a small code action after the webhook to convert some dates, but the it just creates a record.
I did get all of the data into the system but I have about 5 duplicates of each which now means I used about 25000 automation tasks (which should never have happened), but I assume as HubSpot didn’t get a success message it tried again, and again.
I now have to export all of the data and remove all of the duplicates as I do not want to use anymore automation tasks.
What can be done in this situation, and how can I make sure this does not happen again as I thought one of the benefits of using BIKA was that I do not need to use 3rd party automation tools anymore?
First, the 5 duplicate records are likely due to HubSpot’s retry mechanism. According to their docs, HubSpot retries failed webhooks for up to 3 days starting 1 minute after failure, with increasing intervals (max 8 hours between tries). Workflows don’t retry after 4XX errors except for 429s, which respect the Retry-After header. Since BIKA’s automation failed to return a 200 response quickly, HubSpot kept retrying, creating duplicates.
The automation error (Invalid prisma.database.findUnique()...) points to a connection pool timeout. We’ve flagged this to our ops team for optimization ASAP.
Until this is fixed: Have you considered bulk-importing via Excel for large initial datasets? It’s more stable. Once the base data is in, you could enable automations for incremental syncs to reduce pressure on webhooks. Let us know if you need help setting this up!
Thanks for the reply. I understand about the retries and why I got the duplicates, I suppose my question is that when i sent 500 to the webhook there were no issues, but this time the issue started after probably the first 120, is that because there were so many? Unfortunately once HubSpot starts sending the webhooks there is no way to stop it without losing the leads in the flow at that time. Luckily I was able to disable the automation in BIKA, but your systems would still have received the webhook, but just not done anything with it.
I have considered bulk sending but our workflows in HubSpot send a webhook at specific points in the flow, so sending to Excel is almost impossible to do for these. I can modify the flows to split the data into smaller chunks to reduce the load, but can you tell me what size I should be sending?
Will you let me know when the fix is in place, happy to be a tester for you if it helps on this issue.
Not sure what you mean on this “Once the base data is in, you could enable automations for incremental syncs to reduce pressure on webhooks. Let us know if you need help setting this up!” can you help on setting this up or tell me more?
Our engineers checked the logs and found a temporary database connection issue around the 26th, which likely caused those automation errors you saw the next day. You don’t need to adjust your automation steps or split data into chunks—we’ve already tweaked the system to handle this. Feel free to retry syncing from HubSpot via webhooks now.
The incremental sync workaround isn’t needed anymore since we’ve pinned down the root cause (though we’ll still need time for a permanent fix). Give the automation another shot—it should work smoothly this time. Really appreciate your patience, and sorry for the hassle this caused! Let us know how it goes.
PS: Mind sharing your automation steps? Is it just the webhook trigger + script action + create record action? I want to reproduce the same steps to conduct a stress test.
I held off on running another batch of data from HubSpot till today.
I ran one which created about 500 records so I thought it was all good.
Until I ran another of about 2600, I think I got about 1400 before I could not get back into BIKA. When I did get back in I had 7000 new records (a load of duplicates). Can you see what happened in the logs your end as this shouldn’t happen base on your previous message.
Sorry for the late reply! Let me try to sort this out.
When you mentioned “couldn’t get back into BIKA,” does that mean you were logged out mid-process or the system froze? Just wanna clarify!
The 7,000 duplicates might be from HubSpot retries if BIKE didn’t respond quickly. Any chance you can check “Run History” in BIKE Automation, grab screenshots of the errors during that time, and share them?
Teammates scanned logs around July 15th but didn’t catch the same errors as last time. Let’s dig deeper once you share those screenshots.
Everything froze I was not logged out. But, I could not get to the Automation to disable it for about 10 mins.
I have checked the run history and looking through many pages I can only see successes. What i can see is this below, which shows many many processed at exactly that same time (to the second).
I am scared to send anymore into BIKA as it takes me many hours to try to sort the data out once it is in BIKA (so many duplicates). And, I have to wait till HubSpot has finished it’s retries before I can enable the automation again.
Thanks for your help though! My time zone is set to UK (not sure if that is how the time shows in my screenshot)
