fix dashboard revenue isolation and precision

[ChrisEarlAmar]

No description provided.

[sentry]

Bug: The close() method in DatabasePool nullifies self.session_factory without a lock, creating a race condition where in-flight requests can crash during application shutdown.
_{Severity: HIGH}

Suggested Fix

The close() method should acquire the _initialization_lock before nullifying self.engine and self.session_factory. This will ensure that no new sessions can be requested while the pool is being torn down, preventing the race condition.

Prompt for AI Agent

Review the code at the location below. A potential bug has been identified by an AI
agent. Verify if this is a real issue. If it is, propose a fix; if not, explain why it's
not valid.

Location: backend/app/core/database_pool.py#L57-L60

Potential issue: A race condition exists in the `DatabasePool.close()` method. During
application shutdown, `self.session_factory` is set to `None` without acquiring the
`_initialization_lock`. If an in-flight request simultaneously calls
`db_pool.get_session()`, it will find `self.session_factory` is `None` and raise a
`RuntimeError: "Database pool has not been initialized"`. This occurs in deployments
using default uvicorn settings without an explicit `timeout_graceful_shutdown`, causing
requests in progress during a restart to fail instead of completing gracefully.

_{Did we get this right? 👍 / 👎 to inform future reviews.}

[sentry]

Bug: The _async_database_url function silently passes through unrecognized URL schemes, causing create_async_engine to fail if a synchronous driver is configured via DATABASE_URL.
_{Severity: MEDIUM}

Suggested Fix

In _async_database_url, validate the URL scheme. If the scheme is not one of the recognized async-compatible schemes, raise a ValueError with an informative message to prevent the application from starting with an invalid configuration.

Prompt for AI Agent

Review the code at the location below. A potential bug has been identified by an AI
agent. Verify if this is a real issue. If it is, propose a fix; if not, explain why it's
not valid.

Location: backend/app/core/database_pool.py#L17

Potential issue: The `_async_database_url` helper function does not validate the
database URL scheme. If an operator misconfigures the `DATABASE_URL` environment
variable with a synchronous driver (e.g., `postgresql+psycopg2://`), the function passes
it through silently. This causes `create_async_engine()` to raise an
`InvalidRequestError`. While this error is caught during application startup, it
prevents the database pool from initializing. Subsequent API requests that require a
database connection will then fail with an unhandled exception when they try to
initialize the pool again with the invalid URL.

_{Did we get this right? 👍 / 👎 to inform future reviews.}