FlightAware Outage/Stats Missing?

Just checked my FlightAware page and it’s showing no data for today, the 5th at 4PM PDT:


However, when I check my stats, it looks like everything is working fine. I checked my router logs as well and there were no connectivity issues. Are there any issues going on on the FA side? Odd that no data was received for a full hour.

Hi bioyuki

I had a similar occurrence at around 9 or 10am AEST which is about the same as 4pm PDT.

Log outputs for that time show normal operation. :confused: :question:

I don’t understand why the two charts have different times for when it happened too.

The stats system got restarted just after midnight UTC to add some more capacity; in theory that shouldn’t have eaten a whole hour of stats (at most 5 mins) but I guess something went wrong there and that hour will need regenerating.

Hi obj,

Do you have any insight as to why the top chart in my stats. says this happened between 9 and 10am but the lower chart shows between 10 and 11am as shown in my screenshot?

Am I reading them incorrectly?

Thanks

They are both from the same data so one or the other is probably picking the wrong end of the interval.

@carlmortimer,

My graphs are out by an hour too.

Hourly Collection Graphs (EDT) shows the correct info.

Hourly Received Reports (positions | aircraft) is an hour ahead. It current shows be stats for the 10AM EDT hour when it is only 10:21AM EDT.
Maybe one doesn’t cater for Daylight saving time properly.

Jon

Seeing the same dropout at 1800 CDT as well. I thought it was just my connection to my rPi being flakey again.

One of my sites had a drop during the same hour as you guys, but the other site was fine (same location). :confused:

As obj already mentioned, there was a backend stats migration yesterday at 00:00 UTC in order to accomodate better scalability, and that caused an unavoidable one hour glitch for some users. We try to avoid doing that when possible, but the growth in new users has been significant enough that we had to do something to avoid further backlogs and interruptions later.

Got it. Perhaps a sticky about outages/system issues in the future would be useful instead of a bunch of users here trying to debug thinking there was something wrong on their end? Also, just put the messages in a queue until the migration is done next time :wink:

We weren’t expecting an outage, and the “putting messages in a queue” bit was exactly the bit that was being changed, so…