Some worflows just disappeared

I recently installed n8n on docker (on a raspberry pi 4), and I’m very pleased with all the automations I could achieve. Excellent job!
However, I encountered an issue a few times: some of my workflows (sometimes even active and triggered by intervall) just disappeared from the list.
I think it may be related to me duplicating some other workflows and/or saving of a workflow that did not complete (stating there was already a workflow by that name, which was not the case).
Is there a way to investigate?
thanks for your help

Hey @cie172!

I am glad that you like n8n and were able to build automations :slight_smile:

May I know which database are you using (the default is SQLite)? Did you check the database?
Can you also tell if this happens after you restart or update n8n?

You may find this helpful (while inspecting the database):

Thanks for your answer.
Yes, I’m with the default database, and I did not check (was not sure what tool to use, I will check Dbeaver).
No, it did not happen after I restarted or updated n8n.

If it was happening after restarts it could have been a problem with your configuration. Let us know if you find anything.

Whenever you duplicate a workflow, n8n adds the word copy to the name of the workflow. Hence, the new workflow is created with a unique name. When you duplicate a new workflow do you see similar behavior?

Hi there,

I have the same problem.
Yesterday, all my recent workflows and executions history have disappeared. Only data prior to June 15 remains.

I have downloaded the database (sqlite) and browsed it with DBeaver, no trace of recent data.

My server has not returned to a previous backup or anything else, so I’m quite confused.

Do you have any suggestion ?

Hey @SamDsseau,

Welcome to the community :sparkling_heart:

I am sorry that you faced the issue. Did this happen after a recent update or changes that you might have made?

@krynble might be able to help here

Hi @SamDsseau and @cie172

When running n8n in a docker container with default settings, n8n will save all information using SQLite, which is basically a file inside the docker container.

Unless you map this to a folder on your host machine using volumes, whenever a new image is downloaded or a new container is started you will get a fresh new database.

So with this in mind, it is crucial that you run the docker run command with -v ~/.n8n:/home/node/.n8n according to the docs available here: Docker Installation | Docs

I also recommend you backing up this file periodically to avoid losing any information related to your workflows and executions.

I hope these tips help you identifying the problem. If not, feel free to add any more context.

1 Like

Hi @harshil1712,

The only thing that can have operate an issue is a new credential connexion (Microsoft Excel).

During the half hour of the rollback, I was setting up the Microsoft Excel integration.
I would say the disappearance occurred during the minutes I left the Office 365 integration window open while waiting for the callback url to receive the data.

Nothing else happens during this period, except maybe the launch of a workflow by a Cron trigger.

I check with my colleague if it is possible to get back the previous database.

We haven’t found a way to restore the lost workflows.

The issue did not happen on the server side, but could have happened on the Docker or n8n side.

Does anyone have any idea of a solution to recover the previous data?


I just had the same issue. Simply lost a workflow – another that was currently open did not disappear.

We’re running in kubernetes and use a postgres db for the state (without a volume)

Welcome to the community @wftywuf!

Can currently not think about a reason why that would happen. Did you check the Postgres database if they are still saved there and only do not get displayed anymore for some strange reason? Does n8n still behave correctly or is maybe the whole database connection gone?

Generally is there no functionality in n8n which would delete workflows.

Yes, that’s what makes it so weird. Everything else is normal. And the DB was still there. The workflow was deleted from the database. We had a backup, luckily.

I got a modal telling me the workflow was deleted when all this happened. But I was not editing, nor viewing the workflow at that moment. I was testing another workflow. I might have just misclicked – but it would be weird that I could accidentaly delete a worfklow.