feat(streaming): spin loop when too many row-id requests in single ms

[yezizp2012]

What's changed and what's your intention?

As title, after #2205.

Please explain IN DETAIL what the changes are in this PR and why they are needed:

• Summarize your change (mandatory)
• How does this PR work? Need a brief introduction for the changed logic (optional)
• Describe clearly one logical change and avoid lazy messages (optional)
• Describe any limitations of the current code (optional)

Checklist

• I have written necessary docs and comments
• I have added necessary unit tests and integration tests

Refer to a related PR or issue link (optional)

[codecov]

Codecov Report

Merging #2213 (d52b92d) into main (55cfc66) will increase coverage by 0.00%.
The diff coverage is 94.73%.

@@           Coverage Diff           @@
##             main    #2213   +/-   ##
=======================================
  Coverage   70.94%   70.94%           
=======================================
  Files         657      657           
  Lines       83065    83072    +7     
=======================================
+ Hits        58928    58938   +10     
+ Misses      24137    24134    -3     

Flag	Coverage Δ
rust	70.94% <94.73%> (+<0.01%)	⬆️

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

Impacted Files	Coverage Δ
src/source/src/row_id.rs	89.65% <94.73%> (+1.41%)	⬆️
src/common/src/types/ordered_float.rs	24.30% <0.00%> (+0.19%)	⬆️
.../src/executor/managed_state/aggregation/extreme.rs	90.29% <0.00%> (+0.26%)	⬆️

📣 Codecov can now indicate which changes are the most critical in Pull Requests. Learn more

[BugenZhao]

LGTM!

[skyzh]

In fact, I didn't think this really solves the problem. The clock can go forward or backward after the spin loop. So the next time we acquire an epoch, it might still be less than last generated one.

[skyzh]

Okay, maybe this PR doesn't take clock-go-backward into account. Then let's fix this altogether after monolithic clock has been implemented.

[yezizp2012]

Okay, maybe this PR doesn't take clock-go-backward into account. Then let's fix this altogether after monolithic clock has been implemented.

This is just for row-id generate. Yes, clock-go-backward also need to be considered when compute node reboot. This will be fixed together with epoch.