Horizon : properly assess max nb workers permitted by a server (ex: Digital Ocean) to treat a queue of hundreds jobs?

Using a droplet on DO 1vCPU for example and having lets say 500 jobs (http rate limited calls), how we can assess the quantity of worker acceptable for a server?

And using same quantity of worker does it make any difference to split these jobs in different queue compared to keep them is the same queue?

Thks

aurawindsurfing

8 years ago

Level 50

Probably similar to max-childern in PHP-fpm.

Just go to terminal and see how much memory one process consumes, leave some room to breathe and the rest you can use for your tasks. It is just a simple thinking but I remember this is how I was doing it with php children.

https://myshell.co.uk/blog/2012/07/adjusting-child-processes-for-php-fpm-nginx/

Dave97

8 years ago

Level 1

@aurawindsurfing , thanks for the link.

But apparently these command instructions run on my Digital Ocean server doesnt output any results, just a line a header:

S UID PID PPID C PRI NI RSS SZ WCHAN TTY CMD

but no attached content.

Please or to participate in this conversation.