you are trying to run 10 docker browser containers on 3 cpu, every browser container requires 0.5-1 cpu, try to decrease limin. that is why by default is 5...

I have the impression that this option is not working, am I using it correctly after "none" or would I have to use None without "" ?

It's not returning any error, but it still waits for the page to load

"goog:chromeOptions": {"args": ["disable-notifications"
                                                    "disable-infobars",
                                                    "start-maximized",
                                                    "disable-blink-features=AutomationControlled",
                                                    "page_load_strategy=none"],

not directly in capabilities

"pageLoadStrategy": "none" doesn't work as I understood

I didn't understand what you meant! Could you send me an example?

Is there a correct way to use threds with selenoid to avoid possible errors?

Does anyone have examples for me to base myself on?

Selenoid runs browsers in completely isolated containers, so there are no issues with concurrency...

Всем привет, подскажите при попытке создания новой ссессии, контейнер не поднимается и падает с такой ошибкой:


{
   "sessionId": "54de3582dafe7564826aebf1286e6740",
   "status": 13,
   "value": {
       "message": "unknown error: Chrome failed to start: crashed.\n  (unknown error: DevToolsActivePort file doesn't exist)\n  (The process started from chrome location /usr/bin/google-chrome is no longer running, so ChromeDriver is assuming that Chrome has crashed.)\n  (Driver info: chromedriver=89.0.4389.23 (61b08ee2c50024bab004e48d2b1b083cdbdac579-refs/branch-heads/4389@{#294}),platform=Linux 5.10.25-linuxkit x86_64)"
   }
}

или памяти не хватило или какие-то флаги неправильные передаете через капабилити

А есть это m1 )

С этимиже флагами на другом компе все ок

На м1 запускаете? Образ для x86_64...

новая проблема... =)

я так понял нужно запустить на amd64

образы все нужно пересобирать

sorry for offtop question but..... Let's say I am running tests on moon cluster with 15-50 parallel threads. What if those all tests simoultaneuosly start performing selenium wait for condition actions with polling interval set to 100 millis (which means probably lots of wire protocol requests per single time unit and moreover in parallel)? Can this somehow overload the grid hub ?