chore: add debug log for Nacos config snapshot hits

[EvanYao826]

What does this PR do?

Adds a debug-level log message when a Nacos config value is loaded from the local snapshot cache instead of being fetched from the remote server.

This aids troubleshooting in cluster mode where snapshot cache hits indicate a config change notification arrived before the remote node synced.

Note: The core stale-config fix was already handled by #4319. This PR only adds observability for snapshot cache behavior.

[CLAassistant]

All committers have signed the CLA.

[EvanYao826]

I have read the CLA Document and I hereby sign the CLA

[xuxiaowei-com-cn]

I have read the CLA Document and I hereby sign the CLA

• But there might be issues in some places.
• Please sign again, and after signing, click recheck.

[EvanYao826]

Thanks for the heads up! I have rechecked and signed the CLA. Please let me know if there are any further issues.

[xuxiaowei-com-cn]

Thanks for the heads up! I have rechecked and signed the CLA. Please let me know if there are any further issues.

It has been found that you have not signed the CLA. After signing, the CLA will appear green instead of the current yellow.

[EvanYao826]

I have rechecked and signed the CLA. The status should be updated now. Please let me know if there are any other issues. @xuxiaowei-com-cn

[xuxiaowei-com-cn]

I have rechecked and signed the CLA. The status should be updated now. Please let me know if there are any other issues. @xuxiaowei-com-cn

After signing, the CLA will appear green instead of the current yellow.

[EvanYao826]

I have read the CLA Document and I hereby sign the CLA

@xuxiaowei-com-cn I have signed the CLA again. The commit email is correctly set to the noreply format (155432245+EvanYao826@users.noreply.github.com). Could you please check if the CLA status has been updated? If it still shows as not signed, I may need to click the recheck link again.

[xuxiaowei-com-cn]

I have read the CLA Document and I hereby sign the CLA

@xuxiaowei-com-cn I have signed the CLA again. The commit email is correctly set to the noreply format (155432245+EvanYao826@users.noreply.github.com). Could you please check if the CLA status has been updated? If it still shows as not signed, I may need to click the recheck link again.

After you have completed the signing process, please click recheck.

[xuxiaowei-com-cn]

GitHub Actions run has been approved.

[uuuyuqi]

Hi @EvanYao826, thanks for the PR. This overlaps with #4319, which was opened earlier and also fixes #4296, while additionally covering a few related consistency concerns (atomic read-and-remove of the snapshot, distinguishing an empty snapshot from a missing one, namespace-aware key, plus unit tests). We're going to move forward with #4319 to close out the issue, so this PR will be closed in favor of that one.

Before submitting future PRs, a quick scan of open PRs and issue comments helps avoid duplicate work — but the effort here is appreciated regardless.

[EvanYao826]

Rebased onto latest 2025.1.x and resolved merge conflicts. The upstream already merged #4319 with a similar config snapshot fix — my PR now builds on top of that, adding the isEmpty() null-safety check and a debug log for snapshot cache hits.

[EvanYao826]

I have read the CLA Document and I hereby sign the CLA

[EvanYao826]

Hi @uuuyuqi, thanks for the review! You're right that this overlaps with #4319. After rebasing onto the latest 2025.1.x, the upstream already merged #4319's config snapshot fix. My PR now adds incremental improvements on top:

1. isEmpty() null-safety check for the snapshot
2. Debug-level logging for snapshot cache hits

If these additions are useful, I'm happy to keep this PR open. Otherwise, feel free to close it if the team considers #4319 sufficient. Let me know!

[EvanYao826]

Hi, friendly ping — CLA is signed and all CI checks are passing. Could a maintainer take a look when available? Thanks!

[uuuyuqi]

Hi @EvanYao826, I checked the updated diff after the rebase.

The debug log itself is fine, but adding config.isEmpty() is not a null-safety improvement. It changes the behavior intentionally introduced by #4319:

• null means no snapshot exists, so falling back to ConfigService#getConfig(...) is appropriate.
• An empty string means a snapshot exists and the pushed configuration was intentionally cleared. In that case, the empty snapshot must be used without fetching remote config.

If an empty snapshot falls back to a Nacos node that has not yet synchronized, the loader may read the previous non-empty configuration again. This reintroduces the stale-config scenario that the snapshot mechanism is intended to prevent.

The latest CI run also confirms this regression through the following failing test:

NacosConfigDataLoaderTest.loadWhenSnapshotIsEmptyThenDoesNotUseRemoteConfig

Since #4319 already implements the original fix, please remove the isEmpty() fallback. After that, the only remaining change would be the debug log, which would be better handled separately if the project needs it. As currently written, this PR should not be merged.

[EvanYao826]

@uuuyuqi Thanks for the detailed review and explanation! You're absolutely right — the isEmpty() check changes the semantics of #4319's snapshot mechanism. An empty string snapshot should be treated as "config intentionally cleared" and must not fall back to remote config.

I'll remove the isEmpty() fallback immediately. Since #4319 already implements the core fix, the only remaining change in this PR would be the debug log for snapshot cache hits. Would you prefer I:

1. Strip this PR down to just the debug log addition, or
2. Close this PR entirely since fix: use Nacos config snapshot in ConfigData loader #4319 covers the fix?

Working on removing the isEmpty() change now.

[uuuyuqi]

Thanks for the quick update. Please go with option 1 and keep this PR focused on the debug log addition only.

I checked the latest diff, and it now only adds the snapshot-hit debug log, which is the right direction. Please also clean up the commit history and PR metadata so this PR no longer presents itself as the fix for #4296, since that was already handled by #4319.

A single focused commit would be clearer, for example:

chore: add debug log for Nacos config snapshot hits

Please also update the PR title and description accordingly, and remove the Fixes #4296 wording. After that I can take another look.

[EvanYao826]

Hi @uuuyuqi, I've cleaned up this PR as requested:

1. Single commit: chore: add debug log for Nacos config snapshot hits — squashed into one focused commit on top of the latest 2025.1.x.
2. Only the debug log remains: No isEmpty() check, no behavioral change — just the snapshot cache hit log.
3. Updated PR title and description: Removed the Fixes #4296 claim, since fix: use Nacos config snapshot in ConfigData loader #4319 already handles the core fix.

Ready for another look when you have time. Thanks!