Change log level for noisy logs

[thinkharderdev]

Which issue does this PR close?

Closes #.

Rationale for this change

These two logs statements were creating a lot of noise. They can be valuable for debugging but think they should be debug logs instead of info logs.

What changes are included in this PR?

Are there any user-facing changes?

[Ted-Jiang]

LGTM

[mingmwang]

Hi, @Ted-Jiang @yahoNanJing

I think in the pull based task assignment model, if the system is idle, maybe we should let the executor slow down the polling interval. Otherwise there will be too many RPCs between Scheduler and Executor if there are many Executor nodes in the system even when the system is idle.

[Ted-Jiang]

Hi, @Ted-Jiang @yahoNanJing

I think in the pull based task assignment model, if the system is idle, maybe we should let the executor slow down the polling interval. Otherwise there will be too many RPCs between Scheduler and Executor if there are many Executor nodes in the system even when the system is idle.

Thanks @mingmwang its a good idea. Now we hard code it in 1 mins:
https://github.com/apache/arrow-datafusion/blob/8de2a7610df33adf75ba515c4ee2fe935f7aa114/ballista/rust/executor/src/executor_server.rs#L263
How to test the best interval is a problem. @mingmwang ans suggestions 😊?