Historical Maps not updating in hour up to 00:55 UTC
#1
I am seeing one-hour gaps in the history at the same time every day on the Historical Maps. Run the map for 1 hour up to 00:55 and then up to 00:100 UTC on any of the dates below, or the current date. At 00:55, I see no white, yellow or light orange data (sferics within 20, 40 or 60 minutes). At 01:00, sferics for the past hour reappear along with new ones, which is evidence that the data was gathered but was and is somehow held back from the subsystem that underlies the Historical Maps.

It is clear from the regular clock times, calendar dates and global reach that this is some kind of service or developer issue and not a natural phenomenon.

My comment is that many users will be mystified or may misinterpret the data, like dusanfrode (see below), who said, "thunder stopped/slowed down at exact same time all over the world."

My questions are:
1. What was the reason for these data dropouts on the first day of each month, and the periods late October 2016 to late March 2017, and late October 2017 to now?
2. Why were they always at 00:00 (or 23:55) to 00:55 (or 01:00) UTC, and on the 1st of the months?
3. Does the community intend to accept for these to continue indefinitely?

I am a newly registered user. I have searched but found no reference to this specific topic in the Public Forums (one exception below). Apologies if there is an answer in another thread.

Detailed timeline of the occurrences:
The first date I can access the Overview Map is 06:00, 12 November 2015. Since that time, the phenomenon has taken place on the following dates, in the 55-minute or 1-hour period up to 00:55 or occasionally 01:00 UTC, worldwide:
1 July 2016
1 Aug 2016
1 Oct 2016

31 October 2016 to 26 March 2017
1 April 2017
1 May 2017
1 June 2017
1 July 2017
1 August 2017
1 September 2017
1 October 2017
30 Oct 2017 onwards to 29 July 2018 (today).

This was referenced in thread https://forum.blitzortung.org/showthread.php?tid=1672
"The hole [sic] world at the same time" started by dusanofrode.
DelandeC said then, "Unfortunately there has been a network outage in the datacenter yesterday."
Egon provided a GIF showing 2016-04-05 09:00 to 2016-05-05 09:00 UTC, which showed the outage at 2016-04-04 13:25 to 14:25.
The difference was that that was a one-off outage at an arbitrary time of day. Interesting, though, that it seems to have been 1 hour in length.h

read
The hole world at the same timeTh
Reply
#2
I think I found the problem, but I'm not sure yet. At 0:00 UTC there is for one hour no transfer of data from the computing servers to the server we create the images. The data is not lost, only the maps have not been created with the current data. I'll keep looking for the bug, next night at 0:00 UTC Smile

Regards,
Egon
Stations: 753, 762, 1403, 1556, 2022
Reply
#3
Egon, I agree, the data is not lost, just "held back" until 01:00.  I note it happens when the hour resets to zero, and sometimes in the past only when the day-of-month reset to one. Thanks for taking this seriously and getting onto it so soon. Good bug hunting wishes!
Reply
#4
I found the bug this night, it was a simple mistake when creating a directory for the day change, it should work now.

The old maps are not recalculated.

/Egon
Stations: 753, 762, 1403, 1556, 2022
Reply
#5
(2018-08-06, 02:46)Egon Wrote: I found the bug this night, it was a simple mistake when creating a directory for the day change, it should work now.

The old maps are not recalculated.

/Egon

Great, thanks!

On some days at 01:00, for example 1, 2, 4, 10, 18, 25, 26, 28 June, 2 , 3, 14, 21, 23, 24, 27, 30 July and 4 August 2018, data of different ages (20/40/60 minutes) was present/absent from the maps for the different continents/regions of the world. 

It is a credit to the visualisation design with the colour-coded ages of sferics, and thanks to the availability of the dynamic maps, that this issue was so easy to see and that these more detailed variations can also be spotted. 

I look forward to seeing if everything is OK from now onwards.

Andrew
Reply


Possibly Related Threads...
Thread Author Replies Views Last Post
  Historical strike data scjsjo 9 13,482 2015-10-30, 23:40
Last Post: frisch



Users browsing this thread: 1 Guest(s)