test: verify KILL cancels INSERT SELECT

[QuakeWang]

I hereby agree to the terms of the GreptimeDB CLA.

Refer to a related PR or issue link (optional)

Closes #6334

What's changed and what's your intention?

This PR adds tests to verify that tracked INSERT INTO ... SELECT statements can be cancelled through the process manager.

The tests cover both supported kill forms:

• KILL QUERY <connection_id>
• KILL '<frontend>/<process_id>'

The test uses a pending source table scan to keep INSERT INTO target SELECT * FROM source running, then verifies that KILL makes the original insert task return Cancelled instead of only disappearing from SHOW PROCESSLIST.

This does not change runtime logic. It only strengthens coverage for the existing insert-select process tracking and cancellation path.

PR Checklist

Please convert it to a draft if some of the following conditions are not met.

• I have written the necessary rustdoc comments.
• I have added the necessary unit tests and integration tests.
• This PR requires documentation updates.
• API changes are backward compatible.
• Schema or data changes are backward compatible.

[gemini-code-assist]

Code Review

This pull request introduces unit tests to verify that 'KILL QUERY' and 'KILL [process_id]' commands successfully cancel 'INSERT INTO ... SELECT ...' operations. To support these tests, mock components such as 'PendingRecordBatchStream' and 'PendingDataSource' were implemented, and test setup helpers were refactored. Reviewer feedback recommends replacing several instances of '.unwrap()' with proper error propagation and context using 'snafu' to align with the project's error handling standards.

[killme2008]

Good job