Crons not triggering and node retry options disappeared

[d17e.dev]

Describe the problem/error/question

Updated to v1.15.2 a couple of days ago and noticed a couple of regressions?

What is the error message (if any)?

• In my node settings I can’t find the retry options anymore. Is that a scrapped feature?
• Crons don’t seem to be getting triggered consistently. I’m seeing this across different workflows and node types. One workflow with executions that aren’t getting triggered is a Google Drive trigger that checks every minute for new files. Only thing that seems to work no is if I manually trigger the flow. The other flow has a cron that runs every day at 2am and has decided to not do that anymore. At least sometimes…

Is there a way to inspect the state of the cron ‘engine’?

Thanks in advance!

[d17e.dev]

Ok so I found that the missing retry options is a feature:

fix(editor): Hide not supported node options

My question: why does it not make sense to have a retry policy on a trigger node that polls an external service? Several times I’ve seen that the workflow fails because it couldn’t contact that service. In those cases I just want the trigger/poll to run a couple more times before failing the workflow.

For the crons not triggering, this is the trigger node that doesn’t seem to work:

[MutedJam]

@d17e.dev

My question: why does it not make sense to have a retry policy on a trigger node that polls an external service? Several times I’ve seen that the workflow fails because it couldn’t contact that service. In those cases I just want the trigger/poll to run a couple more times before failing the workflow.

Hi @d17e.dev, I am sorry you’re having trouble. For most polling trigger nodes the retry logic is built-in. The Google Drive trigger node would for example update the “last checked” time only at the end of the polling process. So if a polling attempt errors before reaching this line, the missing data should simply come through in the next run.

Now based on your description this does not seem to work for you. Can you confirm how exactly you’ve deployed n8n? Especially which version of n8n you are using and which time zones you have configured for your workflow, for your n8n instance and the value of your TZ environment variable would be helpful here.

[d17e.dev]

Hey! Sorry, forgot to mention my setup:

• n8n version: 1.15.2
• Database Postgres 12
• n8n EXECUTIONS_PROCESS setting (default: own, main):
• Running n8n via Docker on a VPS
• Operating system: VPS on Ubuntu 22.04.3 LTS
• GENERIC_TIMEZONE=Europe/Amsterdam

I had just updated from 1.14.2 where I hadn’t noticed any issues.
I think the problem is specific with the Google Drive trigger no though, as it doesn’t seem to pick up the new files. Other crons I have run as expected. Since the polling executions seem to run silently if there’s nothing new, all I get to see in the execution list is these:

with this as error:

{
"message":"getaddrinfo EAI_AGAIN www.googleapis.com",
"name":"Error",
"stack":"Error: getaddrinfo EAI_AGAIN www.googleapis.com\n    at GetAddrInfoReqWrap.onlookup [as oncomplete] (node:dns:108:26)",
"code":"EAI_AGAIN"
}

The succesful run at the top is manually triggered with the ‘fetch testevent’ in the editor. When I do that, all is fine.

Yet, when the flow is active, new files I put in the folder aren’t picked up. There’s several flows I set up like this with the same ailment.

[MutedJam]

@d17e.dev

getaddrinfo EAI_AGAIN www.googleapis.com

Hi @d17e.dev, thanks for sharing these details! This error is not version-specific but rather suggests a problem with the node’s DNS resolution.

It is, however, not clear to me why new files aren’t picked up for you I am afraid. Can you try reading the TZ variable inside n8n? You can do this by executing a Set node with an expression of {{ $env['TZ'] }} for example:

[d17e.dev]

hmm… looks like you might be on to something here… that variable returns null. Should also be Europe/Amsterdam, as I have this variable in my docker-compose & env files:

[d17e.dev]

Aha! Just found out that the TZ variable sets the system time from the docs pages (Docker | n8n Docs)

so I added that one, and things seem to be actually working now!
Thanks for the help!

[d17e.dev]

Right, I celebrated too soon.
Google Drive trigger still isn’t properly triggering on new files…
Triggering it manually worked well for a little while and now I’m seeing

without me changing anything about the query… I’m not getting the sense that this Google Drive node is very reliable to be honest… Considered building a more robust solution that has simple cron trigger and does an operation for each file in a Google Drive folder, but listing files in a folder doesn’t seem to be straight forward there either, or am I missing something?

[MutedJam]

I am so sorry. I wonder if this could be a permissions problem. Are these files in question owned by you and have you also authenticated the credentials in use by the trigger node?

As a workaround, perhaps you want to consider using the Schedule trigger node along with the regular Google Drive node? I find this approach to be much more transparent and typically prefer it over the polling triggers. In your case, something like this perhaps:

This example workflow runs every hour and fetches items created within the last hour. This way, you fully control the polling logic and can debug every single execution if needed.

[d17e.dev]

Awesome! That’s actually something I attempted to do but I got stuck figuring out the correct query string. Thanks to your little example, I was able to figure it out though. I think I’m also happier with this setup as it allows for a little bit more oversight in the polling process and the possibility to setup retries if necessary.

I’m a lot more confident in these workflows now. Thanks!