[Feature Request] Send a signal to Skyvern's running process to stop the agent which sends webhook #1246
Comments
|
Are you able to retrieve the task I'd when you initially start the task? |
|
@suchintan No, since its only available in the terminal I am running skyvern from. I would have to extract text from the terminal output and regex parse it, which is a lot of overhead. |
|
I see — you're not triggering all of the tasks remotely and relying on the webhook to alert the remote server as to the generated task ids? |
|
Yes that's right. The script creating the task ID's is Skyvern's own. I have no way of accessing the task ID's from the outside unless I actually go to the directory which has the name which is the task ID. |
|
Could you have the script do a webhook call to your server after it's finished running w/ all the task_ids? |
I am running some evaluations currently on the Skyvern web agent. I set a 3 minute timeout, in which I kill the port 8000 the Skyvern process is running in. Once this happens, Skyvern doesn't send a webhook. I need the task ID sent in the webhook response, so that I can copy Skyvern's logs over to my project.
What I would like is to send a signal to Skyvern (such as through an API call) letting Skyvern know that it should halt the currently running process and send a webhook response with the details of that task as was sent in the success cases. This will avoid me having to kill the port and have no webhook response.
The text was updated successfully, but these errors were encountered: