feat(sdk): always print the traceback inside of the wandb.init context manager

[kptkin]

Fixes WB-9886
Fixes WB-10858
Fixes #4225

Description

What does the PR do?

Testing

How was this PR tested?

Checklist

• Include reference to internal ticket "Fixes WB-NNNN" and/or GitHub issue "Fixes #NNNN" (if applicable)
• Ensure PR title compliance with the conventional commits standards

[codecov]

Codecov Report

Merging #4603 (80f592e) into main (df2c7db) will decrease coverage by 0.08%.
Report is 1 commits behind head on main.
The diff coverage is 85.71%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4603      +/-   ##
==========================================
- Coverage   77.87%   77.80%   -0.08%     
==========================================
  Files         379      379              
  Lines       43746    43749       +3     
==========================================
- Hits        34069    34039      -30     
- Misses       9624     9657      +33     
  Partials       53       53              

Flag	Coverage Δ
unittest	81.42% <85.71%> (-0.08%)	⬇️

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

Files Changed	Coverage Δ
wandb/sdk/wandb_run.py	91.03% <85.71%> (-0.23%)	⬇️

... and 14 files with indirect coverage changes

[darabos]

Any chance to merge this tiny change? It's apparently been a problem for two years! (#2387) We just started our W&B trial and hit it on day 3. And the PR is right here! I've patched it in and it works great!

I tried sweeps from a notebook, and I was getting errors only in the sweep but not with normal runs. It looks like sweeps run in a separate process, so I can't use shared data. Which is fine. But this is probably a common way to get sweep-only issues, and they are painful to debug without the stack trace. Thanks for the PR! 👍

[kptkin]

Any chance to merge this tiny change? It's apparently been a problem for two years! (#2387) We just started our W&B trial and hit it on day 3. And the PR is right here! I've patched it in and it works great!

I tried sweeps from a notebook, and I was getting errors only in the sweep but not with normal runs. It looks like sweeps run in a separate process, so I can't use shared data. Which is fine. But this is probably a common way to get sweep-only issues, and they are painful to debug without the stack trace. Thanks for the PR! 👍

After this awesome request, we can't say no. So we will include it in the upcoming release. 😄

[darabos]

After this awesome request, we can't say no. So we will include it in the upcoming release. 😄

Yes!!! Can't wait, thanks!