feat: support pagination on task_history

[ZhiHanZ]

I hereby agree to the terms of the CLA available at: https://docs.databend.com/dev/policies/cla/

Summary

Previously when single request return body is large, it would show grpc error, this Pull Request support the following things:

1. task_name, schedule_time filter push down, allow to use those select information to filter unneeded rows on cloud control services.
2. add next_page_token and page_size parameter, allow to control page size and pagination of the single request.

NOTICE: I also tried to use limit push_down, but found that it does NOT work, need some deep dive on the issue

fix #14929

Tests

• Unit Test
• Logic Test
• Benchmark Test
• No Test - Explain why

Type of change

• Bug Fix (non-breaking change which fixes an issue)
• New Feature (non-breaking change which adds functionality)
• Breaking Change (fix or feature that could cause existing functionality not to work as expected)
• Documentation Update
• Refactoring
• Performance Improvement
• Other (please describe):

[BohuTANG]

NOTICE: I also tried to use limit push_down, but found that it does NOT work, need some deep dive on the issue

We should make it works:) Think a user has many tasks all scheduler is in 1 second, the rows is very large, and run:
select * from system.task_history is no sense anymore.

So, here the better way is:

1. select * from system.task_history returns rows with a LIMIT(Say 10000)
2. select * from system.task_history limit N returns N rows order by the time.

[ZhiHanZ]

NOTICE: I also tried to use limit push_down, but found that it does NOT work, need some deep dive on the issue

We should make it works:) Think a user has many tasks all scheduler is in 1 second, the rows is very large, and run: select * from system.task_history is no sense anymore.

So, here the better way is:

1. select * from system.task_history returns rows with a LIMIT(Say 10000)
2. select * from system.task_history limit N returns N rows order by the time.

first limitation is easy to limit on server side, but for the second part, we need the push_down_info struct to pass limit number to the storage table source, but for now it does not work

https://github.com/datafuselabs/databend/pull/15047/files#diff-d4c61e93b55aeeb6dc42f6eef504cd58c6bdf2e3337d6be71aeeec07c4365973R145-R148

[BohuTANG]

NOTICE: I also tried to use limit push_down, but found that it does NOT work, need some deep dive on the issue

We should make it works:) Think a user has many tasks all scheduler is in 1 second, the rows is very large, and run: select * from system.task_history is no sense anymore.
So, here the better way is:

1. select * from system.task_history returns rows with a LIMIT(Say 10000)
2. select * from system.task_history limit N returns N rows order by the time.

first limitation is easy to limit on server side, but for the second part

This limitation is ok if we give a default number for it.

[hantmac]

LGTM 🚀