You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
While you definitely shouldn't intentionally restart or exit the headless Chrome instance while rendora is working, I agree there should be some way to check the liveness and support automatic reconnections that can be especially useful for failed and restarted headless chrome docker containers. The problem is that chrome-devtools-protocols works basically over websockets and the headless chrome listen websocket url changes every time Chrome is restarted since it contains a UUID.
I suggest we check the liveness every short interval through the endpoint "/json/version", compare the debugging ws urls and renew the ws connection if it is not the same. This might cause unintentional problems with service discovery setups (i.e. the DNS resolution could route you to a different container after each restart).
Since you mentioned puppeteer in the other issue, do you know how is this situation handled there if it's handled at all?
As we had previously used headless Chrome and had an experience that we had to restart the Chrome periodically to make it stable enough, making rendora usable after the headless Chrome got restarted would be important.
When the headless chrome is restarted, rendora will not able to render anymore.
Restart rendora will get this fixed.
The text was updated successfully, but these errors were encountered: