[atlas] Long measurement delays

Hello, Atlas crew,

I’ve been seeing delayed query responses to one-off measurements since November 22, with the exception of November 25.  For www.globaltraceroute.com, which attempts to get responses for 90 seconds after submitting a query (and is usually fairly successful), this has meant a failure rate of 61% on average for those days, and as high as 83% on one of the days.

I haven’t been tracking how long it does take for a response to come in in those cases, but my most recent test measurement, measurement ID 23502037, is still not returning any data an hour and 40 minutes after submission.

My request format  as a Python string is "https://atlas.ripe.net/api/v2/measurements/%s/latest/?key=%s" % (measurement_id, key).  For these failures I’m receiving an empty list, so it looks like this might be something different than the “synchronizing” issue people were complaining about over the last few weeks.  Trying to access that measurement via the Atlas Web UI, if I go to the “latest results” tab I get a spinning progress wheel, seemingly forever.

I haven’t been tracking the Atlas “consumption delay” along with the results, so I’m not sure what it’s looked like during most of the failure periods.  I checked at the time I submitted measurement 23502037 and it showed as green, although it’s showing as yellow now.

Thanks for any help or attention you can provide on this.

-Steve