[enhancement](stats) Param tune which would get a better balance between execution time and cost

[Kikyou1997]

Proposed changes

1. Change the default parallel execution analyze task to 3, which would reduce 15% - 20% cpu cost and increase the execution time a litle
2. Use count(1) - count(col1) rather than sum(case when) which could leverage some optimize like read count from tablet metadata directly

Further comments

If this is a relatively large or complex change, kick off the discussion at dev@doris.apache.org by explaining why you chose the solution you did and what alternatives you considered, etc...

[Kikyou1997]

run buildall

[doris-robot]

(From new machine)TeamCity pipeline, clickbench performance test result:
the sum of best hot time: 47.04 seconds
stream load tsv: 556 seconds loaded 74807831229 Bytes, about 128 MB/s
stream load json: 20 seconds loaded 2358488459 Bytes, about 112 MB/s
stream load orc: 65 seconds loaded 1101869774 Bytes, about 16 MB/s
stream load parquet: 33 seconds loaded 861443392 Bytes, about 24 MB/s
insert into select: 29.1 seconds inserted 10000000 Rows, about 343K ops/s
storage size: 17161955036 Bytes

[englefly]

is there overflow? UNION_ALL_LIMIT is more than 512

[Kikyou1997]

No, in fact we could union even ten thousand query according to my test on legacy optimizer

[morrySnow]

we should move all statistics paramters from config into global variables to avoid restart fe.

[github-actions]

PR approved by at least one committer and no changes requested.

[github-actions]

PR approved by anyone and no changes requested.

[Kikyou1997]

pick from #26858