dashy icon indicating copy to clipboard operation
dashy copied to clipboard

[BUG] Something is going on related to CORS and status checking

Open lyra56k opened this issue 1 year ago • 3 comments

Environment

Self-Hosted (Docker)

System

Rocky Linux 9.4, Podman 4.9.4-rhel

Version

3.1.1

Describe the problem

When enabling status checking, every service shows down with an ETIMEDOUT error. These services are all in fact running. When I manually query the status check endpoint, the same thing happens. This also happens when I go in my Docker container and query 127.0.0.1/status-check with the right headers set. This error happens no matter what I pass to the status-check endpoint. I've tried with and without https://, I've tried sites like Google, I've tried the Dashy instance itself.

Additionally, the cors-proxy endpoint returns ETIMEDOUT for everything in the same manner. I can't use the RSS feed widget with local checking, and I assume this is why, but I'm not sure if that's a separate issue.

No errors show in the Dashy logs, and the errors in my browser console aren't very helpful, they tell me that 'unsafe-inline' doesn't appear in the CSP header, but I very explicitly set that header on my reverse proxy. I don't think that's related though because, like I said, it happens when Dashy queries itself too.

Additional info

No response

Please tick the boxes

lyra56k avatar Sep 13 '24 22:09 lyra56k

I do no use the Dashy status lights. I have 'ETIMEDOUT' timeout errors as well when I turn them on. Mine appears to be because of network configuration on my end. The Dashy container cannot access the network that the DNS provides it to do the healthcheck.

The troubleshooting document seems like it may be helpful in this instance. It has steps I would investigate should I desire this working.

hockwill avatar Sep 14 '24 17:09 hockwill

I actually just checked the DNS of the container and it seems to be working, I can use cURL in there to load the Dashy website itself and everything I want to check the status of. I just took a look at the troubleshooting document and it doesn't really seem to answer this question unfortunately...

lyra56k avatar Sep 14 '24 18:09 lyra56k

For anyone who might stumble across this issue, you pretty much have to set the network_mode to host in Docker. I don't at all like this solution because now I'm forced to allocate port 8080 to the container, but it is what it is. At least the status checks work now I guess...

lyra56k avatar Apr 17 '25 21:04 lyra56k