EF Core default log limits for migration failure diagnosis

What EF Core logs by default for a migration failure?

When a migration throws an exception, EF Core writes an Error entry in the Microsoft.EntityFrameworkCore.Infrastructure category. The log contains:

• The exception message (e.g., Invalid column name 'Foo')
• The migration name that caused the failure
• A stack trace pointing to IMigrationsAssembly.ApplyMigrationAsync
• Any DbUpdateException details (e.g., SqlException code)

It does not include the exact SQL that was executed or the parameter values used in the command.

Effect of EnableSensitiveDataLogging()

Enabling sensitive‑data logging changes the Microsoft.EntityFrameworkCore.Database.Command category only:

• Parameters are appended to the command text in the log (e.g., INSERT INTO Users (Name) VALUES (@p0) becomes @p0 = 'Alice').
• The migration error log itself already contains the exception message; the exception stack trace is unchanged. If the exception contains parameter values (e.g., a DbUpdateException that wraps a SqlException), those values are now visible because the underlying SqlException is logged with the full command text.

In short, EnableSensitiveDataLogging does not alter the error‑level log for migration failures, but it enriches command‑level logs and can surface parameter values inside the exception payload.

DbContext Pooling and Log Correlation

When AddDbContextPool is used, the same DbContext instance can be reused across requests. Log scopes (e.g., BeginScope with an RequestId) are attached to the DbContext and therefore can persist across multiple operations. This means:

• Startup logs and migration logs may share the same scope identifier, making it harder to isolate a single migration run.
• Correlation IDs are not automatically reset between requests; you need to create a new scope for each operation or rely on a unique identifier per migration execution.

For diagnostic clarity, consider adding a custom scope around migration code:

using var scope = logger.BeginScope("MigrationRun:{Id}", Guid.NewGuid());
context.Database.Migrate();

This guarantees that each migration attempt has a distinct log correlation ID, regardless of pooling.

Practical Verification Steps

1. Configure logging to Debug or Trace in appsettings.json or via ILoggerFactory.

2. Run a migration that will fail (e.g., add a column with a missing type). Observe the Error entry under Microsoft.EntityFrameworkCore.Infrastructure.

3. Repeat with optionsBuilder.EnableSensitiveDataLogging() and verify that command logs now include parameter values.

4. Switch to AddDbContextPool and add a custom scope around the migration as shown above. Check that the Id differs between runs.

What’s Missing?

To fine‑tune your diagnostics, it would help to know the exact exception message EF Core produced for the failed migration. If you can share that message, we can confirm whether the stack trace includes the SQL command and whether sensitive‑data logging is indeed exposing the parameters.