*: improve some logs

[rleungx]

What problem does this PR solve?

We recently encountered a problem, which PD cannot delete the leader key due to somewhere hanging. So the new leader cannot be elected. Such problems are not easy to trace once the scene didn't be saved. We can't check the goroutine status. Under this circumstance, it's hard for us to know if this hanging is caused by our code or etcd. For now, we need to wait for some goroutines to exit before deleting the key. And we don't know if these goroutines exit accurately.

What is changed and how it works?

This PR improves some logs to help us trace the problem better.

Check List

Tests

• Unit test

[nolouch]

/rebuild

[codecov-io]

Codecov Report

Merging #1511 into master will increase coverage by 0.19%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #1511      +/-   ##
==========================================
+ Coverage   67.48%   67.68%   +0.19%     
==========================================
  Files         158      158              
  Lines       15550    15557       +7     
==========================================
+ Hits        10494    10529      +35     
+ Misses       4094     4074      -20     
+ Partials      962      954       -8

Impacted Files	Coverage Δ
server/region_syncer/server.go	81.89% <100%> (ø)	⬆️
server/cluster.go	75.26% <100%> (-0.35%)	⬇️
server/leader.go	78.83% <100%> (-1.42%)	⬇️
server/coordinator.go	84.4% <100%> (+0.06%)	⬆️
client/client.go	68.77% <0%> (+0.21%)	⬆️
server/handler.go	55.58% <0%> (+0.51%)	⬆️
server/schedule/operator_controller.go	93.65% <0%> (+0.79%)	⬆️
server/schedule/region_scatterer.go	88.23% <0%> (+2.2%)	⬆️
... and 6 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 22032cb...7cc2821. Read the comment docs.