I’m collecting data for a project where I expect to collect sample sizes of up to 500-1000 participants across different conditions. I’ve managed to build a stand-alone Flask application for my experiment. I have a load balancer (nginx), and webserver (Werkzeug) all set up. Can anyone suggestion an appropriate server configuration? i.e. How much RAM is ideal? Using a SQLite database should I be concerned about storage? How many CPU cores do people often use when hosting?
My current set up is throwing server overload errors so I’m thinking I need to change something.
hi @Bryan_Gonzalez , welcome to the forums! I hope someone might be able to help you answer your question, but just a heads-up that anecdotally, many people on this forum use services such as Qualtrics, Pavlovia, or Gorilla to collect data (so there might not be a whole lot of expertise around the server configuration settings). @Josh - is there someone on the Prolific team with this sort of expertise that could provide some advice here?
I’m afraid this is a little outside our wheelhouse. I’ve asked the wider community. Let’s see if someone else can help! Sorry, that we can’t be of much use here.
As Josh says, your questions are a bit beyond the scope of the Prolific community to date at least.
The question as to how to set up servers so as to cope with hundreds of simultaneous users is one which comes up quite a lot in relation to online education. These sort of questions (number of processors, amount of ram, load balancing) come up quite a lot on the Moodle forums and there are presentations
e.g. Moodle Clustering: How to Support Hundreds of Users - unstac.tk…
explaining how it is done. Moodle is one platform you might think of using since it has survey modules.