In T278611#6951218, @RhinosF1 wrote:

Can you ask them? There's a big difference between years and today. Not sure what in the last 24 hours could have caused it given yesterday should have been no deploys.

Not sure, but I think it's been going on for years, many months at least. I just never bothered to report the issue until now.

I have the same problem.

It might also be worth considering to only show median values:

• If there are irregularities, the median is the more robust value and should be shown instead of the mean.
• If there are no large irregularities the mean and the median shouldn't differ much and either one could be shown.

The implementation would then be relatively simple by swapping the mean calculations for a call to a median function and adapting the documentation accordingly. Unfortunately I have no experience in node development so won't be of much help with the actual work.

In the last couple of days I regularly encountered this error with varying hosts (s4.web..., s5.web..., ...). Reloading the page sometimes works, sometimes it throws the same error again.

This was probably caused by commit 4436724a57176e4111550f186e297006ec924320 several hours ago.
Replacing

if (!$clusterNr) {

with