ci: use the shared scheduled-failure report

[lesnik512]

Switches the scheduled-failure report to the shared workflow, per the updated standard CI2 (modern-python/.github#127).

Changes

• scheduled.yml's report-failure job now calls modern-python/.github/.github/workflows/report-scheduled-failure.yml@main, the job CI2 requires byte for byte.
• .github/scripts/report-scheduled-failure.sh is deleted. CI2 forbids a per-repo copy.

The shared workflow keeps the label, issue title and open-or-comment behaviour. Only the issue text changes: it no longer lists likely culprit packages, since the per-repo lists had drifted. It points at the linked run instead.

Verification

• actionlint 1.7.12 passes on scheduled.yml.
• The shared workflow was tested end to end from its branch in eof-fixer before docs: align README idle-latency figure to docs (~10ms) #127 merged: the first run opened an issue, and the second commented on it. See feat: share the scheduled-failure report as a reusable workflow .github#127.
• The new job only runs on a scheduled failure, so this PR's CI can't exercise it.

[github-actions]

Benchmark gate

✅ gate passed

scenario	msg/s	delete/msg	WALrec/msg	WALB/msg	fpi	upd	del	dead_tup
consumer/w1/b10	1277	1.000	6.80	916	242	5000	5000	10000
consumer/w1/b100	1204	1.000	6.84	919	243	5000	5000	10000
consumer/w2/b10	1482	1.000	6.79	958	242	5000	5000	10000
consumer/w2/b100	1886	1.000	6.73	957	243	5000	5000	10000
consumer/w4/b10	1433	1.000	6.77	960	242	5000	5000	10000
consumer/w4/b100	2215	1.000	6.87	1035	243	5000	5000	10000
consumer/w1/b100/tfbs100	8844	0.010	6.13	1157	265	5000	5000	10000
producer/w1/b100	3198	0.000	3.04	584	0	0	0	0

Gated (fails the build): delete_calls + tuple counters (upd/del/ins) + the producer's insert_calls, exact; select_calls within +2; wal_records within a 10% band. msg/s, WAL bytes and total calls are informational (timing/FPI noise).