-
-
Save benuski/f60c973424cb7c061ac103ea1a02ca96 to your computer and use it in GitHub Desktop.
| version: '3.2' | |
| services: | |
| changedetection: | |
| image: ghcr.io/dgtlmoon/changedetection.io:latest | |
| container_name: changedetection | |
| hostname: changedetection | |
| volumes: | |
| - changedetection-data:/datastore | |
| environment: | |
| - PORT=5000 | |
| - PUID=1000 | |
| - PGID=1000 | |
| # Ensure this URL matches the name of the playwright service | |
| - PLAYWRIGHT_DRIVER_URL=ws://playwright-chrome:3000/?headless=false | |
| ports: | |
| - 5000:5000 | |
| restart: unless-stopped | |
| depends_on: | |
| - playwright-chrome | |
| playwright-chrome: | |
| hostname: playwright-chrome | |
| image: ghcr.io/browserless/chrome | |
| restart: unless-stopped | |
| environment: | |
| - SCREEN_WIDTH=1920 | |
| - SCREEN_HEIGHT=1024 | |
| - SCREEN_DEPTH=16 | |
| - ENABLE_DEBUGGER=false | |
| # Increased connection timeout to reduce chances of timeout errors | |
| - TIMEOUT=600000 # Now 10 minutes | |
| # Increased concurrent sessions for better parallel processing | |
| - CONCURRENT=15 | |
| volumes: | |
| changedetection-data: |
Hi. Thanks for this, does this not work for sites using javascript?
Upd: Here's the docker-compose.yml I used. Works great. Most sites run without a proxy, only a few require one. No warnings or errors.
https://pastebin.com/pATyMgu0
I also used a proxies.json as described in the manual here: https://github.com/dgtlmoon/changedetection.io/wiki/Proxy-configuration (you can leave this out if you don't use proxies or prefer to set them in the web UI instead).
Edit: Updated the docker-compose.yml. This version has been running without any issues for over a year now.
- If you're running it behind a reverse proxy, comment out the
ports:section and adjust the networks accordingly. - Browserless chromium is the way to go. I tried every other compatible browser and this is the only one that works reliably with browser steps and the visual filter.
- You may want to tune
shm_size,FETCH_WORKERS,andCONCURRENTto your needs. It runs fine on a cheap VPS, but expect around 500 MB RAM idle. If you have a beefier server, feel free to crank those up. - Replace
your-very-secure-tokenwith something randomly generated. I used "openssl rand -hex 32". BASE_URLis only relevant if you set up notifications and want the correct server name displayed there, so don't worry too much about it.- The
HEALTH,MAX_CPU_PERCENT,andMAX_MEMORY_PERCENTsettings are there so browserless doesn't crash under load. Instead of dying, it'll return an error and keep running. Useful on a low-end VPS.
Thank you. It is working for most pages I want to monitor, but not for https://www.smythstoys.com/at/de-at
I think the problem is Imperva. Do you know any solution to that problem?
What error are you getting? If it's only one website, I'd recommend setting up a proxy for that specific domain. The site likely doesn't like your IP.
I use Bright Data's datacenter proxies for this. There's an official guide for setting that up. I had similar issues where certain sites wouldn't work without a proxy and this solved it.
If that doesn't work, they also have a Browser API, though it's expensive. I'd only use that for sites that absolutely require it.
Error: Access Denied.
The Docker instances are running on my NAS with my local IP. From my PC I can access the page without any problems.
It's likely Imperva detecting the scraping pattern rather than blocking your IP specifically.
Try going to Settings → Fetching and set Random jitter seconds ± to 10 (or any value other than default) to add variability between requests.
If that doesn't resolve it, proxies are your only solution.
Shame that I can't add an emoji to a gist but same here, super helpful (and the comments) thanks!
I'd love an example of how to leverage browserless.io's unblock as part of changedetection too