Folder Quikstart Quide?
I feel like this is a very common use case: You have a list of scripts in a directory and you want to be able to run them with different schedules.
In my case I have a python scripts folder with "requirements.txt" but I can't seem to figure out what the easiest way to deploy this is. Theres so many fancy features but getting something like this up and running just doesnt seem possible? I dont want to edit one script at a time. I'd rather have everything synced to git and then push once in a while. Folder push seems to expect a file???
How do I do this? I feel like what I want to do is not that unique....
InternalError: Sql error: error communicating with database
Heya, I'm facing an issue in scheduled flow runs where the database seems to be unavailable sometimes, which leads to the schedule getting deactivated. Any idea how I can prevent this?

Python pycache
We have quite big internal library, bootime for a worker is around 20sec. anything to improve?
Unable to build docker image
Hello all,
I'm trying to build the docker image from the docker file at
I'm constantly running into dependency build issues. Below is the error:
```error: could not compile idna (lib) due to 1 previous error
36.80 warning: build failed, waiting for other jobs to finish......
windmill/Dockerfile
windmill/Dockerfile
Unexpected Const Error
When running a typescript (bun) step I receive this error:
534 | const settled = results
error: Unexpected const
at input.tsx:534:3...
Question on worker memory usage
We have configured a worker group in our compose file with a hard limit of 1024M. However, we still see successful flow executions with a peek much larger - is this expected behavior? I suspect this is Docker deciding to not kill the process if it exceeds temporarily the memory limit, but that we should not rely on this behavior being predictable & instead scale up our worker limit to be a safe % above the peak we observe. Is that assumption correct or do we miss something in the way this works?

Execute a script for each flow run
Is there a way to run a specific script for each flow run? We would like to do things like custom logging, without having to update each & every individual flow.
For example, is it possible for the worker to have a "post flow run" hook for example that can be called?...
Worker caching weirdness
We are using a web KV store. For a specific key we have written three blobs of JSON - each write with a different number of array elements in this order (5, 8, and 16). We have a simple windmill typescript script that fetches the data and logs the number of array elements. When we run that script on the windmill deploy (self-hosted EE) the workers get a variety of responses from the fetch (sometimes 5, someimtes 8, sometimes 16). Its like windmill has cached the internal fetch results and is returning that instead of doing the actual fetch. Which seems weird to me - since its just running our script on a worker. However, if I run the exact same fetch on a cloud-hosted windmill, the script always gets 16 elements (which is all that instance has ever seen). Is WM doing something like this or have we just hit some weird distributed caching problem with the KV store we're using?
The fetch is simple:
let result = await fetch(url)...
How to differentiate between preview and production runs?
As the title says: how do I know inside of my script if it is running in preview or production mode?
I noticed that I might be able to use env variables for this (in the settings - runtime menu), and this works if you are just running the script standalone. But, when I wish to run a flow in preview mode, the scripts inside of the flow do not seem to realize they are running in preview mode, and the production environment variables are still used....
puppeteer is failed to start
I had a working browser automation script and from last week it's not working
so I created a new script using this guide https://www.windmill.dev/docs/advanced/browser_automation (bun puppteer version) and still it's not working
Auto scroll took 30929ms mem peak: 79.43MB
...
Custom url for public app
Hey, we are creating some public apps using windmill and would like to assign a custom url to these pages.
For example, I want a user to be able to browse to windmill.company.com/my_app and be shown my page, without them seeing a redirect to /public/foo/bar in their browser. This would make it way easier for them to bookmark stuff.
Is this possible through the caddy files? We couldn't get it to work....
Integrating an app in a custom React application
Is there a certain design concept we can follow (and is it technically supported) to integrate a Windmill app in a different (react) application, and pass along things like authorization cookies / information to render data in apps specific to the context of a user outside of Windmill? We are now using a combination of public & private apps, but want to explore the option for more fine-grained access control based on authentication information outside of the Windmill App context. Thanks!
Can't get custom React component to show up on screen
I've followed the guidelines here: https://www.windmill.dev/docs/apps/react_components
But the component which I'm supposed to start editing doesn't seem to load correctly on the browser. However, when I build it and upload it to windmill, it works flawlessly. Any help here is welcome! Thanks!
Attached is the error I get as of right now....

Support for HTTP Browser Redirects in Windmill Functions (302 Response Handling)
Hey guys,
I'm trying to implement URL redirects through a Windmill TypeScript function.
Specifically, I want to create an endpoint that, when accessed through a browser, redirects the user to another URL (in my test case, to Google.com)....
Python requirement issues
Hey,
I am trying to write a script that makes use of the azure opentelemetry sdk.
When I run the script below locally, it just works. But when running in windmill. I get this error:
ModuleNotFoundError: No module named 'azure.monitor.opentelemetry'
...Quota exceeded: Email trigger usage limit of 100 per day has been reached
We are receiving "Quota exceeded: Email trigger usage limit of 100 per day has been reached" for one of our flows. However, we're only having 10-20 inbound emails per day, throughout the instance. Is there any overview over all runs that have been triggered via email trigger so we can investigate this further?

port used for resume URLs
Hi, I've written a approval script for my flows, when generating resume urls with
wmill.get_resume_urls()
, it keeps generating urls for port 8080
even though my windmill server is sitting on the default port 80
. Therefore, whenever I try to use the generated resume urls, it fails and I am forced to manually switch the url to use port 80.
For dev, I've just got the windmill server up and running using docker compose up
. Is there any other config I need to do to get resume URLs working?...Branch All Parallel throwing Error
In this configuration, which is simple echo bash scripts at present throws error mentioned in the image. I tried to check in GitHub issues and documentation about "Branch", but found no instructions about what could be the reason for this error.
Edit #1: Error shows up after "Shared Directory" is enabled.
Edit #2: Solved. ...

Windmill on Windows without a Docker
Hello,
I noticed that Windmill supports native Windows functionality without Docker only in the Enterprise Edition. Are there any plans to release free-to-use binaries for Windows in the Community Edition?
I’m interested in exploring Windmill for local hobby projects at home but would prefer not to use Docker or virtual machines. Hypervisors tend to consume significant resources, even with minimal configurations like a single VM with 1 vCPU and 2GB RAM in an idle state. This makes Docker and VMs inefficient options on Windows for lightweight use cases....