Webhook gives a success response, but workflow is not executed until almost a month (!) later

Questions
1

Describe the problem/error/question

We are on the cloud-hosted version with a pro account. So far we have been quite happy with the product. During the first week of testing and then going live everything ran super smooth.

What is the error message (if any)?

I just checked our executions and where it should say around 50-100 per day, on the 13th of October it says 2116 successful executions (in a single day!).

Since we store some data as execution data, I was able to find out the following (we send “order created” events for an online shop and set the customer e-mail as execution data, so I looked at one of the strange executions).
We call the webhook and log success (2xx response) or failure (non-2xx response). All our calls are logged as successful, and this is done using a standard http client (java unirest), which has given us 0 issues over the last 20 years.
For this e-mail we see:

Webhook called on Sept 18 → workflow executed
Webhook called on Sept 22 → workflow executed
Webhook called on Oct 1st → No workflow executed
… 6 more between Oct 1st and 13th → no workflow executed

And then all of a sudden the missing workflows are executed started at around 15:48 on Oct 13th. This cannot be a caching issue on our end, there is no proxy, we did app restarts in between etc.

An example of a workflow that we triggered on Oct 1st at 07:17 (CEST) which ran on Oct 13 at 15:48:03:

We call the webhook when the order is created, which, as can be seen in the webhook data was at 07:17 (“createdAt:2025-10-01T07:17:02”).

What gives?

It appears that the webhook calls somehow get (got?) stuck inside the n8n cloud and did not get flushed until Oct 13.

This, of course, makes the integration unreliable. I hope someone at n8n can shed a light on this for us. We are a paying customer (pro account, prepaid for a year).

Thank you for your attention to this matter.

Please share your workflow

This is not related to the content of the workflow.

Share the output returned by the last node

Also not relevant to this issue.

2

There is something else that is off. Under insights, we see this:

n8n-insights
n8n-insights2396Ă—1390 184 KB

But in the admin panel we see a different number altogether:

n8n-admin-panel
n8n-admin-panel348Ă—204 3.62 KB

So it appears that the admin panel does count the executions correctly (based on submission time?), but the insights is misreporting or something like that.

3

Hi Team, it’s a bit scary that there has been 0 response to my request. How are we supposed to get any support? Again, we are a paying customer. Thank you.

4

Can you verify the following:

The date ranges in both views

Any active filters in the Admin Panel

Your plan’s execution retention limits(execution logs are automatically pruned based on your plan (e.g., Start/Starter plans keep max 2500 executions with 7 days retention). The Admin Panel might be showing pruned data while Insights retains aggregated metrics.)

5

Hi! Thank you for your response.

Both screenshots were taken at the same moment.
There is no filters we can activate in the admin panel. On the insights screen you can see it is set to the “Last 7 days”.

Since the screenshots were taken no more than 5 minutes apart I doubt a filter is the issue.

…. and that is more of a secondary issue.

My main concern is that many of our executions were stalled for up to two weeks. The reason I pinged this topic is that the team may be able to have a look at logs now, but not for an unlimited time due to log retention on your end.

1 Like
6

After receiving an automated message asking me if this has been solved: no, this has not been solved yet.

1 Like
7

@Jon_James or @robtf9 maybe can help in this… sry :disappointed_face: , can’t help further.

8

Hi @bgooren

You’re a Pro cloud user, and on October 13th, you saw a massive, unexpected spike of over 2,000 workflow executions, far more than your usual 50-100 per day.

After investigating, you found that webhooks your system successfully sent between October 1st and 13th didn’t trigger any workflows at the time. Instead, all these delayed webhooks were suddenly processed in a large batch on October 13th.

This makes your integration unreliable, and you are seeking an explanation from the n8n team for why the webhooks got stuck and executed much later.