known keyring to determine the key and fingerprint. So the older the
upload, the more unlikely it is that the key is still in the keyring:
udd=> select extract (year from date) as year,
count(*),
count(case when fingerprint = 'N/A' then 1 end),
count(case when signed_by = 'N/A' then 1 end)
from upload_history
group by 1
order by 1 desc;
year | count | count | count
------+-------+-------+-------
2024 | 9835 | 3488 | 2687
2023 | 45189 | 728 | 728
2022 | 57792 | 1110 | 1088
2021 | 44760 | 333 | 333
2020 | 51791 | 1891 | 1891
2019 | 39710 | 3617 | 3616
2018 | 47934 | 4101 | 4101
2017 | 34879 | 2909 | 2909
2016 | 46990 | 3931 | 3931
2015 | 36589 | 2780 | 2780
2014 | 35260 | 8992 | 8992
2013 | 31344 | 14045 | 14045
2012 | 31469 | 15832 | 15832
2011 | 32882 | 18602 | 18601
2010 | 29099 | 22663 | 22622
2009 | 32542 | 31200 | 31171
2008 | 31996 | 31662 | 31662
2007 | 28315 | 28109 | 28102
2006 | 32650 | 32551 | 32550
2005 | 27461 | 27308 | 27303
2004 | 30057 | 29709 | 29654
2003 | 27440 | 27391 | 27389
2002 | 27411 | 27411 | 27411
2001 | 24983 | 24983 | 24983
2000 | 10860 | 10860 | 10860
1999 | 12977 | 12977 | 12977
1998 | 3102 | 3102 | 3102
(27 rows)
This would probably need to be re-thought.
There's also something strange in 2024 that needs to be investigated:
udd=> select changed_by_email, count(*) from upload_history where fingerprint='N/A' and date > '2024-01-01' group by changed_by_email order by 2 desc limit 10;
changed_by_email | count
----------------------+-------
vorlon@debian.org | 414
bdrung@debian.org | 240
tille@debian.org | 130
jbicha@ubuntu.com | 104
zigo@debian.org | 97
werdahias@riseup.net | 97
mwhudson@debian.org | 84
slyon@debian.org | 79
ginggs@debian.org | 54
dr@jones.dk | 53
(10 rows)
Lucas