How to debug the execution which takes a long time unusual?

Questions
, ,
1

I have a cron job workflow that is very simple:

  1. HTTP Request to get data
  2. Function node to parse result (to one item)
  3. HTTP Request to post the parsed data to another API

Normally, it only takes 0.5 seconds to 3 seconds, sometimes if the data is larger like 3000KB, it will take 14s to finish the whole workflow.

But today I found an execution, the data was only 500KB, and it takes 68 minutes to finish, I check the API, no timeout(response immediately).

Are there any ways to check where the execution stuck?

2

Hey @petersiow, welcome to the community! I am sorry to hear you’re having trouble.

But today I found an execution, the data was only 500KB, and it takes 68 minutes to finish, I check the API, no timeout(response immediately).

Can you confirm where you got the two values (immediately + 68 minutes) from?

On a general note, when debugging things my first step would be to enable debug logging (described here) and then taking a look at the server logs for the respective time range.

Some possible causes for a slow n8n instance could be a high load on either the server itself or a large/slow database. For the latter you could consider implementing a pruning strategy as described here. Be careful though, when using SQLite and enabling DB_SQLITE_VACUUM_ON_STARTUP, your first n8n startup will take a very long time - so you might want to do this outside of your core business hours.

3

Thanks @MutedJam

I check the executions log (on the n8n panel), and it shows the running time is 68 minutes.

About the “response immediately”: I checked the timestamps of the cronjob start time and the access log of the API server (steps 1&3). Both of them are very close (diff only a few seconds). It should be able to prove that there is no timeout or stuck. Therefore I think all the nodes work correctly and get the result immediately.

I agree that the large/ slow database is one of the reasons to cause this. Due to our n8n having many cronjobs being run simultaneously and in high frequency, it keeps so many execution logs.

I will try to prune the execution log, refine the schedule of the tasks, and see if it solves the problem.

Thank you :smiley: