Re: Improving tracking/processing of buildfarm test failures

Hello hackers,

Please take a look at the February report on buildfarm failures:
# SELECT br, COUNT(*) FROM failures WHERE dt >= '2026-02-01' AND
 dt < '2026-03-01' GROUP BY br;
REL_14_STABLE: 22
REL_15_STABLE: 13
REL_16_STABLE: 3
REL_17_STABLE: 12
REL_18_STABLE: 56
master: 48
-- Total: 154
(Counting test failures only, excluding indent-check, Configure, Build
errors, also excluding failures from fruitcrow.)

# SELECT COUNT(*) FROM (SELECT DISTINCT issue_link FROM failures WHERE
 dt >= '2026-02-01' AND dt < '2026-03-01');
14

# SELECT issue_link, COUNT(*) FROM failures WHERE dt >= '2026-02-01' AND
 dt < '2026-03-01' GROUP BY issue_link ORDER BY 2 DESC LIMIT 6;
https://www.postgresql.org/message-id/739010.1772142042%40sss.pgh.pa.us : 44
-- Fixed

https://www.postgresql.org/message-id/CAOYmi%2B%3D4QhCjssfNEoZVK8LPtWxnfkwT5p-PAeoxtG9gpNjqOQ%40mail.gmail.com : 16
-- Fixed

https://www.postgresql.org/message-id/CA+hUKGL0bikWSC2XW-zUgFWNVEpD_gEWXndi2PE5tWqmApkpZQ@mail.gmail.com : 7

https://www.postgresql.org/message-id/4ccf62c4-48ed-47cb-badc-9ae436d91b39%40app.fastmail.com#6434c9ffe3c955bfbefb5a3748bc7d2f
: 7

https://www.postgresql.org/message-id/9ffdb19a-7a89-424e-925a-dd981c37f0ba%40gmail.com : 4

https://www.postgresql.org/message-id/aPAl27_urHSODwRN%40paquier.xyz : 4

# SELECT COUNT(*) FROM failures WHERE dt >= '2026-02-01' AND
 dt < '2026-03-01' AND issue_link IS NULL; -- Unsorted/unhelpful failures
61
40 of them ("could not create semaphores", timeouts) were produced by
icarus -- looks like it's not ready to bring anything useful yet -- I will
exclude it from consideration for now.

Short-lived failures: 60

Best regards,
Alexander