Pavlovia servers malfunctioning at the moment

Hi,

I am running an online study on Pavlovia, have tested heaps of participants so far without major issues. However, in the last hour participants suddenly started sending me messages that they came to the end of the task and got this error:

“when uploading participant’s results for experiment: milan.andrejevic/tbdtask
504 Gateway Time-out
nginx
Try to run the experiment again. If the error persists, contact the experiment designer.”

When I try to download the data I also get the following error:

"Error when preparing the download of the experiment results:
“malformed server response: missing downloadToken”

If the error persists, contact Pavlovia’s administrator."

This is a pretty big problem right now, as participants are unable to complete the rest of their session.

I would appreciate an urgent response!

Kind regards,
Milan

One of the worst things about this problem is that the default error message seems to be instructing participants to do the task again, I just had a person report they ran my 40 minute task twice because of that error, just to get the same error again!

I am Also having problems with the server, cannot upload a new experiment I want to upload.

It should be working now.

Hi,

it’s still not working (I am still having trouble downloading the data). I am also having trouble finding an appropriate way to contact administrators to report urgent incidents like this - is this forum the only way to report this?
The incident has incurred lots of costs at my end, so it would be great if I could get in touch with administrators via email and sort out the causes and discuss how we can prevent this from happening again.

Best wishes,
Milan

Hello @milan , @Hubertrober ,

I am afraid that we came under very heavy load this morning, owing to somewhat naughty users clogging the pipes.
I intervened and the situation was resolved under a couple of hours.
I am very sorry to read that it impacted the running of your experiments.

@milan Are you still experiencing difficulties? As far as I can ascertain the system is behaving in a nominal fashion since noon UK.

Alain

Hi Alain,

I checked again and I am still not able to download the results, the dashboard seems to be stuck showing a pop up message: “preparing the results for download…” and the download buttons are greyed out.

It would also be great if you or any other admin could respond to my previous message or get in touch with me privately via email.
Also I raised about the default error message that participants receive when this happens.

Kind regards,
Milan

That is rather odd. What is your username? I’ll have a look for you.

milan.andrejevic

The download was stuck in limbo. I’ve fixed it.
You should be able to download your results now.

Thank you for flagging the issue with the error message. We are working on giving experiment designers the possibility of crafting their own messages, since they know better than us what would or should happen in case of difficulties with their participants.

Alain

Hi Alain,

I’ve just worked for 6 hours straight just trying to manage as much as I can the awful consequences of your server malfunctioning during one of my batch testing sessions, and the costs are still very high.

Could you please let me know if someone will be in touch to discuss this with me? Or do I need to voice all my concerns and ask all my questions in public?

As for the error message issue, I am really not sure how it is appropriate in any circumstance to ask participants to do the task again and again while the servers are down so it does not strike me as a customization issue. A much better default error message would acknowledge that pavlovia servers are malfunctioning and that they should contact the experimenters to tell them that.

You describe this as “difficulties with participants” but that isn’t quite accurate. The difficulties are in that this is a participant testing platform that apparently cannot handle the traffic volume it sells, and that doesn’t know how to inform participants of its failings.

Kind regards,
Milan

I am also experiencing the same issue downloading the results after participants reported a 504 Gateway Time-out error. Please would you be able to advise on how I can access the data? My username is ExPrag_UoM. Thanks!

Had the same problem with my tasks on Pavlovia, unfortunately had to pause data collection. Seems to be quicker at responding today hoping to open up collection again. Are you still running into the 504 error? Thanks in advance

I don’t know whether the 504 error is still occurring - I haven’t run any more participants as I’m still unable to download my data. @apitiot please would you be able to release the download?

Good morning,

I have been keeping a close eye on the server performance since Monday, to make sure that there was no further slow downs, and I have been implementing additional measures to prevent further abuse. It seems to me that the performance has been nominal since then.
I am sorry to read that you were negatively affected.

Could you both share the names of your experiments? I’ll take care of the download problem.
Cheers,

Alain

Thanks Alain, much appreciated - mine is at ExPrag_UoM/ae2

You are most welcome.
I have just fixed the issue. You should be able to download your results.
Best wishes,

Alain

I was able to download my data, but my participants were seeing 504 errors with connecting to the tasks as of Wednesday, not Monday. We have a series of 4 tasks and utilizing the router, participants are directed from one task to the next. Participants were unable to go through to the next tasks as they received this time out error. I want to know if it is safe to run my participants through Pavlovia or will they be timed out again. Thanks!

Our names of the experiments with the account Tricam :
Add-It-Up
Episodic_Memory
Sensory_Precision
Working Memory

I would think that it should be fine. I believe I managed to take care of the naughtiness.
Cheers,

Alain

I am also having this issue. My U/N is PaulStott. The experiment is: PaulStott/phd_experimentthreedeepshallowcomp. I’d really appreciate it if you could look into this and see if you can resolve it.