Hanging check

[Noki]

Describe the bug

I have multiple checks setup, and only one fetch_worker due to limited ressources on my NAS. The worker got stuck, so all my checks stopped working:

Version
v0.45.23

To Reproduce
No idea how it happened but the check uses the Playwright Chromium/Javascript

Expected behavior
I would expect the following:

• The check that got stuck should fail with a timeout. There should be the option to set the timeout globally and for specific checks.
• There should be a notification for checks that failed with a timeout
• There should be a notification for checks that could not be run within their usual schedule, so the admin could take actions

[dgtlmoon]

heya, no logs from the container? platform? docker? pypi install? downloaded directly?

[Noki]

The relevant part of the docker log (bottom to top):

Looks like it was detected that the browser steps were incomplete. After that it hang. The other containers did not output any useful log at that time. Hope this helps.

[dgtlmoon]

Looks like it was detected that the browser steps were incomplete.

No its just saying its using a different driver, its just a debug message really

After that it hang.

The logs only show to step 4 (press enter)

any more info?

what about proxies? or external browsers? you didnt say

[Noki]

Step 4 is the last step before evaluation. It submits a form, than the check should take place on the result displayed. I configured a 10 second wait time before extracting the text:

I don't use proxies / external browsers. Here is my docker compose file:

version: '3.2'
services:
    changedetection:
      image: ghcr.io/dgtlmoon/changedetection.io:latest
      container_name: changedetection
      hostname: changedetection
      volumes:
        - /volume1/docker/changedetection/datastore:/datastore

      environment:
        - PLAYWRIGHT_DRIVER_URL=ws://playwright-chrome:3000
        - HIDE_REFERER=true
        - FETCH_WORKERS=1

      ports:
        - 3003:5000
      restart: unless-stopped

      depends_on:
        playwright-chrome:
          condition: service_started

    playwright-chrome:
        hostname: playwright-chrome
        image: dgtlmoon/sockpuppetbrowser:latest
        cap_add:
            - SYS_ADMIN
        restart: unless-stopped
        environment:
            - SCREEN_WIDTH=1920
            - SCREEN_HEIGHT=1080
            - SCREEN_DEPTH=16
            - MAX_CONCURRENT_CHROME_PROCESSES=2

[dgtlmoon]

is it reproducable?

[Noki]

No. Works again after I restarted all containers, therefore some detection would be nice since no checks ran over the weekend due to missing alerting.

[luckman212]

I am hitting this too over at #1836 (comment). Also maybe related: #2293

Running v0.45.23 on a Synology NAS. What kind of logs would be helpful here, if any?

[calvinbui]

im having the same issue, and I use https://github.com/dgtlmoon/sockpuppetbrowser.

is there a way to add a timeout to either the sockpuppetbrowser or changedetectionio?