fix: honor configured query retries and stop retry-event crash

[joshmarkovic]

What

Fixes two defects in the query-level retry path in SQLServerConnectionManager.add_query.

1. The retry-debug event crashed instead of retrying. On a retryable error, the code fired AdapterEventDebug(message=...), but the event field is base_msg. Constructing the event with message= raises a protobuf ParseError, so the first retryable error crashed at event construction rather than triggering a retry. dbt-core's base add_query uses base_msg=; this matches it.

2. Configured retry counts of 1-3 were silently ignored. The retry limit was computed as credentials.retries if credentials.retries > 3 else retry_limit, where retry_limit is a hardcoded 2. So any configured retries value of 1, 2, or 3 was discarded in favor of 2, meaning even the default retries: 3 only attempted a query twice. This now passes credentials.retries directly, mirroring open(), which already drives connection retries off credentials.retries.

Behavior change

With these fixes, query retries actually function (previously the first retryable error raised a ParseError), and the configured retries value is honored. With the default retries: 3, a retryable query is now attempted up to 3 times rather than 2.

Tests

Adds unit tests for add_query's retry path:

• retry until success on a retryable error,
• the retries: 3 regression (attempts the configured number of times),
• no retry when retries: 1,
• non-retryable errors are not retried.

Verification

• Full unit suite: 272 passed.
• ruff, black (line-length 99, py310), and mypy all clean on the changed files.
• dbt/adapters/sqlserver/sqlserver_connections.py was unchanged between the base this work started from and current master, so the change applies cleanly with no overlap against recent work.

[axellpadilla]

Just flagged this for final release 1.10.1