Improve Misleading Error Message in ContentTooLongException

[raynigon]

Fixes #1238

Context

Repository: elastic/elasticsearch-java
Pull Request: #1239
Component: Content validation / Buffer management

Description

When an incoming data stream exceeds the configured maximum buffer limit, the client throws a ContentTooLongException. However, the error message incorrectly reports the size of the current incoming data chunk (src.limit()) instead of the total accumulated content length.

This results in a highly confusing error message for users and developers, as the reported "too long" content size often appears to be significantly smaller than the configured buffer limit, masking the actual reason the request failed.

Example Scenario

Configured Buffer Limit: 1,000,000 bytes

Current Buffer Length: 995,000 bytes

Incoming Chunk (src.limit()): 8,192 bytes

Actual Combined Length: 1,003,192 bytes (Exceeds limit)

Resulting Exception Message: > "entity content is too long [8192] for the configured buffer limit [1000000]"

This makes debugging difficult, as it implies a chunk of ~8KB somehow breached a ~1MB limit.
Expected Behavior

The exception message should report the total calculated content length that caused the threshold to be breached, making it entirely clear why the data stream was rejected.

Expected Exception Message:

"entity content is too long [1003192] for the configured buffer limit [1000000]"

Actual Behavior

The exception message only references the size of the final byte buffer chunk (src.limit()), completely failing to account for the data already stored in the buffer (buffer.length()).

[cla-checker-service]

💚 CLA has been signed

[l-trotta]

nice catch, thank you! LGTM