This comment has been deleted.

It is
https://www.codewars.com/kata/59c53183aeb2842585000116

08-03-1646 21:20
Etc/GMT-2 (+2)
Jamaica (-5)
expected:<08-03-1646 14:[12]> but was:<08-03-1646 14:[20]>

30-04-1875 13:56
Asia/Novosibirsk (+7)
Asia/Khandyga (+9)
expected:<30-04-1875 1[7:2]6> but was:<30-04-1875 1[5:5]6>

The random tests are using ancient time periods which requires arcane knowledge about all the documented time zone and clock changes, which is, obviously, library-dependent, hence implicitly forcing a specific library to be used in the solution.

04-08-2703 18:21
Europe/Simferopol (3.0)
Europe/Warsaw (1.0)

Also, testing times in far future is even more dangerous: we do not know the possible timezone changes in the future, hence even the currently used library will be suspect (unless it can keep updating the time adjustments till eternity).

the input format of date and time is "dd-MM-yyyy HH:mm", 23h clock

There's no such thing as "23h clock".

1. return "Cuckoo Cuckoo Cuckoo Cuckoo Cuckoo Cuckoo Cuckoo Cuckoo Cuckoo Cuckoo Cuckoo Cuckoo" - you are serios? Why not same as "Cuckoo " * 12
2. Why to call time.split(":") twice?
3. Why each if-statement in one line, but the last in three?
   Too much questions.. :D

Is there issue or we sould resolve it ?

You made my day with this comment.

Nice Kata for the Java8 time API.

Actual date time is: your answer should be: 27-03-2021 21:41 expected:<[27-03-2021 21:41]> but was:<[your answer]>

There's no need to show the values twice, showing the input would be more useful.

that does ring a bell, in Java, yeah...

Yes, I know, I thought the same.

I expected this to be a duplicate, but couldn't find any similar kata.

I have the same story.