web-arena-x
waiting for get_by_label("Username or email")
Hi,
I copied AMI in our AWS server and checked that we can visit the virtual website. However, when I tried to reset the environment for specific task, I encountered following error:
playwright._impl._api_types.TimeoutError: Timeout 10000ms exceeded.
=========================== logs ===========================
waiting for get_by_label("Username or email")
============================================================
Do you know the reason for this error and resolvation?
Thank you.
Hi, can you provide the full error history? Seems like the program tries to access an element in a specific page, but the element does not exist.
Thanks for the reply. The above error was resolved by retrying the reset. However, It seems environment is slightly unstable, as error like below occurs randomly during evaluation on 812 tasks.
Traceback (most recent call last): [66/1904]
File "/home/dongjun/autoprompt/browsergym/src/eval_rephrase.py", line 489, in <module>
raise e
File "/home/dongjun/autoprompt/browsergym/src/eval_rephrase.py", line 483, in <module>
obs, info = env.reset()
File "/home/dongjun/anaconda3/envs/webarena/lib/python3.10/site-packages/gymnasium/wrappers/order_enforcing.py", line 61, in reset
return self.env.reset(**kwargs)
File "/home/dongjun/anaconda3/envs/webarena/lib/python3.10/site-packages/gymnasium/wrappers/env_checker.py", line 59, in reset
return self.env.reset(**kwargs)
File "/home/dongjun/anaconda3/envs/webarena/lib/python3.10/site-packages/browsergym/core/env.py", line 255, in reset
goal, task_info = self.task.setup(page=self.page)
File "/home/dongjun/anaconda3/envs/webarena/lib/python3.10/site-packages/browsergym/webarena/task.py", line 103, in setup
self.webarena_instance.ui_login(site=site, page=page)
File "/home/dongjun/anaconda3/envs/webarena/lib/python3.10/site-packages/browsergym/webarena/instance.py", line 89, in ui_login
page.goto(f"{url}/customer/account/login/")
File "/home/dongjun/anaconda3/envs/webarena/lib/python3.10/site-packages/playwright/sync_api/_generated.py", line 9221, in goto
self._sync(
File "/home/dongjun/anaconda3/envs/webarena/lib/python3.10/site-packages/playwright/_impl/_sync_base.py", line 104, in _sync
return task.result()
File "/home/dongjun/anaconda3/envs/webarena/lib/python3.10/site-packages/playwright/_impl/_page.py", line 495, in goto
return await self._main_frame.goto(**locals_to_params(locals()))
File "/home/dongjun/anaconda3/envs/webarena/lib/python3.10/site-packages/playwright/_impl/_frame.py", line 147, in goto
await self._channel.send("goto", locals_to_params(locals()))
File "/home/dongjun/anaconda3/envs/webarena/lib/python3.10/site-packages/playwright/_impl/_connection.py", line 61, in send
return await self._connection.wrap_api_call(
File "/home/dongjun/anaconda3/envs/webarena/lib/python3.10/site-packages/playwright/_impl/_connection.py", line 461, in wrap_api_call
return await cb()
File "/home/dongjun/anaconda3/envs/webarena/lib/python3.10/site-packages/playwright/_impl/_connection.py", line 96, in inner_send
result = next(iter(done)).result()
playwright._impl._api_types.TimeoutError: Timeout 10000ms exceeded.
=========================== logs ===========================
navigating to "http://ec2-18-189-57-2.us-east-2.compute.amazonaws.com:7770/customer/account/login/", waiting until "load"
When I faced such error, It is sometimes resolved by just retrying the run. I am using my own amazon machine by copying the webarena AMI. Have you exprienced such instability?
Thanks for sharing more information. This is a rare error during my experiments. Usually, this error indicates that the website is down. We recommend t3a.xlarge, 1000GB EBS root volume. Smaller instances may result in website crashes.
You can try to access the website from a browser and see if you can still load the pages.
Thanks, we are also using t3a.xlarge, 1000GB EBS root volume instance as recommended in the repo.
Current location of our server is South Korea and I wonder if physical distance between aws instance and our server can be a reason for this issue.
- If it is, one possible solution might be copying the AMI in us-east-2 region to the Asia Pacific (Seoul) region, and intialize the instance. But when I try to copy the AMI, I encountered following error:
You do not have permission to access the storage of this ami
