Please implement support of generating OpenAPI 3.0 self-describing specification.

Flutter is strongest growth repo on Github for mobile development so would be nice that more features from the SS client are implemented

Encryption-at-Rest is an important feature for GDPR regulation compliance.
EF and SqlServer offers an ProtectedPersonalData Annotation to protect fields on db with encryption.
It would be very useful to have a similar feature on Ormlite.

Usage example: https://github.com/blowdart/AspNetCoreIdentityEncryption

Please add support for Streams as you did for Strings, Lists and Sets

Would be very helpful to be able to stream RequestLogger data to AWS Kinesis Data Streams and Kinesis Data Firehose.

The ServiceStack.Aws.DynamoDb has only limited APIs that are async (currently 3 methods).

I would like to see the rest of the APIs get an async overload.

I hope that's enough of a description.

Saving classes with properties of type PhysicalAddress does not work. The reason should be the serialization:

    [Test]

    public void JsvStringSerializerWorksForPhysicalAddress()

    {

        var address = PhysicalAddress.Parse("0123456789AB");


        var serializer = new JsvStringSerializer();

        var result = serializer.SerializeToString(address);

        StringAssert.Contains("0123456789AB", result);

    }

Currently we have to write a testmethod for each endpoint and define the DTOs and then provide data to test. Instead if we have a testfile generated which can act like template and teams just write data against it. This helps teams to have proper test coverage and would save the time to create these request and response models.(Which already ServicestackVS reference does)

The OrmLiteConnectionFactory can register several database servers, as described here:
https://stackoverflow.com/a/13693254/178143

However, to use the servers that is no "default", then you have to do it manually by using the name.

I would suggest having a setting where the database server used is either random, round-robin or some other way. Sure, its not rocket science perhaps to write a wrapper for that yourself, but maybe it would be neat to have support in the framework.

Thank you.

Trying to UpdateOnly, when onlyField is typed as SqlExpression and the query has a Join an error will be thrown that says "The multi-part identifier "DepartmentJobCode.DepartmentId" could not be bound."

Example:
var jobsQuery = _db.From<dev.ScheduleEvent>()

            .Join<dev.DepartmentJobCode>((se,jc)=> se.Job == jc.JobTitleId)

            .Where<dev.DepartmentJobCode>(e => Sql.In(e.DepartmentId, departments))

            .Update(x => new { x.Published, x.ModifiedBy});

db.UpdateOnly(new dev.ScheduleEvent { Published = false, ModifiedBy = userId}, onlyFields: jobsQuery);

1h
https://aws.amazon.com/blogs/aws/new-amazon-dynamodb-transactions/

Add Transactions to IPocoDynamo

When an error occurs processing a batch of requests, the responses for the successful requests are lost. It would be helpful if these successful response are returned along with the error that stopped further processing.

AWS Charges by the IOp, not supporting UPSERT effectively doubles the cost (and reduces the performance) of running mysql/mariadb on that platform with the SELECT/INSERT behavior of OrmLite

I'd like to be able to incorporate my Cognito User Pool as an authentication and role provider.

See issue: https://forums.servicestack.net/t/should-csv-respect-dto-datamember-order-value/1799

The AutoQuery plugin is great especially for use with many front end grids/tables, however it lacks the ability to Group By. Can you implement it to have feature parity to work with many of the front end grid components (ie kendo grid/devexpress/etc)?

Two Factor Authentication is a pretty standard thing nowadays, it even comes baked into the typical ASP.NET Core Identity system. It would be nice if Service Stack had this feature as well. We are working on porting a legacy application from ASP.NET to ServiceStack and this is a requirement for government applications (some form of TFA).

Storage Queues are cheaper and more performant than azure service bus if guaranteed first-in-first-out, etc is not required.

They do offer 'At-Least-Once' guaranteed delivery, where Redis is 'fire and forget'.

Combining this with full featured CosmoDb caching instead of Redis would make a compelling low-cost option.

While some methods of SqlExpression<T> supports up to 15 generic parameters (Where, And, Or) others like Select, SelectDistinct and GroupBy only supports a couple. When building complex queries, it's easy to select data from more than 7 tables, or group by fields from more than 4 tables.

Could you implement more Select, SelectDistinct and GroupBy overloads like you did for Where, And and Or?

There is no specific number of parameters needed, but I think that the default 15 is a good start.