Workflows that fail without warning or error log

[Ragsthenos]

Hello everyone,
I’ve searched the forum for solutions but haven’t found anything, so I’m taking the liberty of asking the question in case someone else has experienced this.

I have workflows that fail randomly on various activities. It can range from a simple query to a basic exclusion activity without any technical parameters.
In short, my workflows can fail at any time and at any activity. The problem is that I don’t get any error messages in the logs, so I don’t know how to proceed.

I have already restarted the entire instance and all the services to no avail.
No network issues detected.

Do you have any suggestions I could explore?
We’re planning to perform a PostgreSQL update today, but I doubt it will resolve the issue.

This issue affects all the workflows in my marketing instance.

[ccg1706]

Hi @Ragsthenos, 

It can be from database performance issues, network or timeout settings , insufficient system resources , or corrupted imput data. You can start optimising queries, validating data, checking workflow logs and reasure that your instance is updated with the latest version.

I leave you with some useful documentation links: Data base monitoring  / Workflow best practices 

Regards, 

Celia

[Ragsthenos]

Hi @ccg1706 

Thank you for the documentation
My instance is in the latest version and my workflows worked and still work, it's just that at times (and it's quite frequent) one of them fails

Regards,

[_Manoj_Kumar_]

Hello @Ragsthenos  When you say they fail what does that means? Are they stuck in processing an activity in a WF or goes in failed/paused state?

If you are not seeing any errors in workflow journal then I am assuming it because of some deadlocks in database. I would suggest you to breakdown queries in small pieces and then see which query is taking long time to execute and then optimize that query by either creating indexes or changing the conditions.

You can also look at control panel to see which query is taking long time to execute.

[Ragsthenos]

Hi @_Manoj_Kumar_ 

Hi Kunal,

When workflows fail, I mean that they are in a failed/paused state.

We also thought about a deadlock but the workflows were working before and the "crash" can happen at any time in any workflow. It can block at the level of a javascript code, a query, a simple exclusion box etc... The activity in question becomes flashing red but there is no log. When I restart the workflow, it manages to execute correctly.

We have enabled Verbose and tracefilter and even with that we do not find anything.

Do you have any other pites?

Kind regards,