fix(controller): collect pod log before delete when cancel job

[jialeicui]

Description

• Collect the pod log before deleting the job
• Use pod name as log file name (remove the timestamp)

Modules

• UI
• Controller
• Agent
• Client
• Python-SDK
• Others

Checklist

• run code format and lint check
• add unit test
• add necessary doc

[codecov]

Codecov Report

Merging #2326 (0addb43) into main (334938d) will decrease coverage by 8.09%.
The diff coverage is 79.16%.

@@             Coverage Diff              @@
##               main    #2326      +/-   ##
============================================
- Coverage     82.69%   74.60%   -8.09%     
+ Complexity     2418     2415       -3     
============================================
  Files           427      328      -99     
  Lines         22664    11546   -11118     
  Branches       1315     1302      -13     
============================================
- Hits          18741     8614   -10127     
+ Misses         3335     2343     -992     
- Partials        588      589       +1     

Flag	Coverage Δ
console	?
controller	74.60% <79.16%> (-0.06%)	⬇️
standalone	?
unittests	?

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
...starwhale/mlops/schedule/k8s/K8sTaskScheduler.java	66.25% <33.33%> (-0.21%)	⬇️
...n/task/status/watchers/TaskWatcherForSchedule.java	95.65% <92.30%> (ø)
...s/watchers/log/CancellableTaskLogK8sCollector.java	64.28% <100.00%> (ø)
.../task/status/watchers/log/TaskLogK8sCollector.java	67.85% <100.00%> (ø)
...ava/ai/starwhale/mlops/schedule/k8s/K8sClient.java	52.33% <100.00%> (ø)
.../starwhale/mlops/schedule/k8s/PodEventHandler.java	59.67% <100.00%> (+0.66%)	⬆️

... and 102 files with indirect coverage changes

[goldenxinxing]

LGTM