Quick question for freelancers and agency folks who deliver n8n workflows to clients.
After you hand over a workflow to a client (whether it’s on their Cloud instance or their self-hosted server) — how do you find out when something breaks?
- Does the client call/message you?
- Do you keep admin access and check manually?
- Do you have some monitoring setup?
And honestly — how much time per month do you spend on “it stopped working” support calls that you didn’t expect?
Asking because I’m trying to understand if this is a pain others face or just me.
1 Like
Welcome to the n8n community @Jawad_Nisar
Usually a mix of all three. I try not to rely only on the client noticing, so the safest setup is to keep some kind of monitoring in place, whether that is error notifications, periodic health checks, or admin access for support. In practice, if there is no monitoring, you usually only hear about it when the client complains, and that turns into reactive support very quickly.
Hi @Jawad_Nisar Welcome!
First you have clear it with the client although if you made the workflow very professionally the client would know automatically via any error branch of a node that would notify the client, i would highly recommend to always tell the client at the project delivery time that “this would not break, if it ever did due to some external service, the flow would notify you & and you can always contact me” this builds a lot of trust not just for one off projects but the client now knows who to pay to fix their system.
No NEVER!! Always make sure you do not keep any of the access after workflow delivery as this can get you in potential trouble of data leak or anything related to that, if client says it then its fine but never keep yourself anywhere in the position of monitoring, that can backfire really quick.
After delivery just tell the client about the scenario of error cases and how the workflow would know the client that there is some error, and what i would recommend is that by using the N8N API make a small dashboard that would showcase kind of a health of all the workflow runs, i have did that for a client and they were really happy that they can always check if there is something wrong in previous runs.
Huf, personally never faced MANY calls as i use a safer and longer approach for building client flows that is like very likely to break and even if it did after delivery and it is not in the contract terms i would make sure to charge the client accordingly to the amount of days passed after delivery if it is less then ofc dont charge if it is more than you can charge no problem there.
I get you @Jawad_Nisar but if you build client flows very professionally the client side rarely breaks and even if it did make sure you setup a lot of extras they would love it!
One concrete setup that works well: n8n has a built-in Error Workflow feature (under Settings → Error Workflows) where you configure a single workflow to trigger whenever any other workflow fails. We use it to send an instant Slack notification with the workflow name and execution ID — that way you hear about failures immediately instead of waiting for the client to call.