Public farm data collection temporarily suspended
Dear PCIbex users,
As you may have noticed, we have been facing repeated outages on the public farm at https://farm.pcibex.net/ in the last few days (the UPenn farm is not affected by this). In order to assess the situation more thoroughly, we have decided that it will unfortunately be necessary to suspend data collection for a few days - the exact timing is hard to predict, but we hope to having things back up soon, and will keep you posted through this list.
We realize that there are many aspects of users interacting with the farm beyond data collection, especially in terms of project development and testing out code. We are therefore not taking the farm completely offline (and it is back online now), but be prepared for intermittent and unannounced outages due to maintenance and update operations from our end. Existing results files will also remain accessible. So the main concrete change in suspending data collection will be that the data-collection links will be temporarily deactivated across the board for all experiments. That way, any participants from recruitment platforms that still try to access an experiment will not be able to start taking it, avoiding unnecessary hassle from participants going through an entire experiment but faltering at the ‘send results’ section at the end (and there data not being recorded). However, it will of course be the best practice for everyone in the meantime to pause data collection on recruitment platforms, so that no participants try to take an experiment but can’t load it.
For anyone that is under urgent time-pressure to collect data, there are some alternative options in the meantime: first, you can move your project to another farm instance (e.g., the previous PCIbex farm or the original IBEX farm), as you can still download the entire project; this is relatively straightforward for simple projects, but if you have resources (e.g., audio or image files) are loading zip-files, or are collecting audio- or video-recordings, more adjustments may be necessary (e.g., host your resources externally, and update your server settings for the source of requests to access zip-files/upload permissions for recordings etc. to the farm you are using). Note that if you have access to a web-server and control over all the relevant settings and configurations, you can also download a stand-alone version of your experiment and collect data on your own server.
We understand this may create difficulties and are very sorry for the inconvenience. The popularity and corresponding usage levels of the new farm are giving rise to new issues, especially in relation to the intricate configurations on the AWS platform, that we weren’t facing until usage peaks of various sorts started happening, and could thus not be foreseen. We hope to have these issues ironed out soon, and will keep users up to date of any developments either way.
Jeremy & Florian