Some ADS-B stats improvements going out today (Jan 19)


#1

We’re rolling out some internal changes to the ADS-B stats processing code on our server-side today which we are expecting to let us make the data on the webpages update more regularly and frequently. There might be some slight disruptions in data updates for a little bit while we finish deploying the new code to our server cluster. These changes should also hopefully reduce some of the confusion when users try to manually update their feeder’s exact location via the website. I’ll post another note in this thread once we believe everything is complete.

Update: we’re still re-processing today’s data (Jan 19), so today will appear with lower numbers until that is finished. Additionally, very new users that joined today might not be visible until that process finishes. We’ll also be re-processing the data for Jan 16 since that day wasn’t migrated properly.


New stats with hourly log
Register multiple sites
Missing statistics graphs
Anyone else's daily stats just drop?
#2

Seems they have had a negative effect. My positions reported dropped from over 60,000 to around 4,000. :cry: Will the lost data be recovered?


#3

My numbers for 16th Jan and 19th Jan are way off (missing loads of positions and aircraft). I guess thats all part of the works going on and they hopefully will be back tomorrow.


#4

I just noticed the 16th being way off before I saw your post. :open_mouth:


#5

Looks like the 16th is a migration problem – we’re going to reprocess that day.


#6

Looks like something exciting happened between 2100 and 2200 UTC 2015/01/19 – that’s when my numbers reset.

I’ve been responsible for always-on, 24x7, low latency transaction systems in the past – now my son does that kind of stuff and I can sit back and laugh. I appreciate what you’re doing and thank you for doing it so well.

cheers–

bob k6rtm


#7

Thanks for the update.


#8

Thanks for the update. I put a new feeder online today and was thinking that somehow I’d managed to blow up the whole system… :wink:


#9

my stats for sat 17th are not right either dropped to around 30% of what they should be


#10

My guess is they are re-processing the 17th as well my stats for the 16th were incorrect earlier today but now are correct.


#11

dunno about it, but only in chome i see some stuff, in the delayed life tracker…


#12

I seem to be missing most of yesterday 19th. I usually clock up about 260,000 positions and 3500 aircraft each day. My yesterday figures are 26,000 and 350 :open_mouth: :open_mouth:

EDIT - mine are sorted now, thanks


#13

I’m missing alot of yesterdays (Jan 19th) stats too.
It should be above 4000 planes detected but it only has 550-ish

I noticed yesterday evening (23:30 CET) that stats for the 16th were wrong too. This seems to be right now.
Stats for the 19th are still wrong. Are those lost or does it need to be recalculated just like the 16th?

One detail i also noticed yesterday though, was the stats for the 19th were building up. probably during processing. I decided to go to bed at midnight and the count for the planes was around 700 for the 19th at that point. Its odd its back to a lower amount now.

EDIT: Seems its being calculated right now, its at 1900 planes now (12:57 CET)


#14

My stats for 19th Jan seem incomplete. When I looked earlier today, they showed 600 planes but now it’s down to 514. Actually, that increased to 516 in the last minute. Perhaps a re-import or migration is running right now. Will keep watching :slight_smile:


#15

I believe all of the data should be caught up now. We have a known issue with newly added receivers not being assigned a location at all, which we are still working on.

mgunther, are you still lower than expected?


#16

All numbers look fine now, thank you very much!


#17

Looking good here. Thanks!


#18

My “Coverage Distribution” graph has changed since the stats upgrade. I had 26398 hits in the <50nm NW segment a couple of days ago. Today I only have 23649. Can you take a look please?

EDIT I’ve just realised that the coverage distribution is a daily graph. DOH !!

I’d be interested in a cumulative graph in a future release.
Thanks


#19

My “Coverage distribution” used to update daily (reset every 24h) but since the enhancements were implemented, it has been showing relatively same state. Positions seem to change but it doesn’t reset to 0 after midnight GMT anymore. Have you noticed the same?


#20

The hits for <50nm in the NW segment have *fallen *to 17693 from 23649 few hours ago.

I haven’t taken much notice when the graph gets zeroed. I’ll keep an eye out.