Skip DDB Enhanced Client extension processing for unusued extensions

[RanVaknin]

[SURFACE API CHANGE - NEEDS REVIEW]

Context

The extension system creates context objects and invokes all configured extensions on every operation. By default, the Enhanced Client includes VersionedRecordExtension and AtomicCounterExtension, which run even when beans don't use either extension.

ChainExtension invokes each extension's beforeWrite() and afterRead() methods, creating a context object for each invocation. With 2 default extensions, this overhead accounts for ~2-15% CPU time, including extension invocation, context creation, and result merging.

This PR adds shouldProcess() method to the DynamoDbEnhancedClientExtension interface with a default implementation returning true for backwards compatibility.
Extensions can override this method to check if they're configured for a specific table schema.

Flamegrahps

Put SMALL

Before
two calls ~8% + ~3% around ~11%

After

Results:

Operation	Size	V2 Master Mean (us/op)	optimized	Improvement
Delete	TINY	0.396	0.431	-8.12%
Delete	SMALL	0.387	0.387	0.00%
Delete	HUGE	0.392	0.384	2.08%
Delete	HUGE_FLAT	0.397	0.39	1.79%
Get	TINY	0.462	0.489	-5.52%
Get	SMALL	0.975	0.98	-0.51%
Get	HUGE	10.198	9.821	3.84%
Get	HUGE_FLAT	3.024	3.056	-1.05%
Put	TINY	0.688	0.625	10.08%
Put	SMALL	1.345	1.22	10.25%
Put	HUGE	13.482	13.107	2.86%
Put	HUGE_FLAT	7.595	7.964	-4.63%
Query	TINY	2.557	2.487	2.81%
Query	SMALL	4.162	4.084	1.91%
Query	HUGE	31.065	29.631	4.84%
Query	HUGE_FLAT	8.975	9.165	-2.07%
Scan	TINY	1.080	1.082	-0.18%
Scan	SMALL	2.577	2.671	-3.52%
Scan	HUGE	30.366	30.25	0.38%
Scan	HUGE_FLAT	8.105	8.357	-3.02%
Update	TINY	1.277	1.259	1.43%
Update	SMALL	9.241	9.305	-0.69%
Update	HUGE	39.486	38.944	1.39%
Update	HUGE_FLAT	232.677	243.128	-4.30%

[RanVaknin]

Results are inconclusive. The overhead we are eliminating is likely cancelled by the new interface method checking.

[github-actions]

This pull request has been closed and the conversation has been locked. Comments on closed PRs are hard for our team to see. If you need more assistance, please open a new issue that references this one.