Integration

we can do this in a Staggered way using GraphQL. Below presents different ways to use the Archiecture.

The Client can use GRPC with GraphQL inside. So Can use Centrifugo.

This can ALL run in a single Process !! Finally https://github.com/systemd/portable-walkthrough-go

A

Simple where:

the DB holds all data for Write and Reads.
We get mutations via DB trigger

How ?

Mine the DB

Get all Tables
Create trigers, and so dont need to do any CDC
Use Gorm for this ?

GraphQL

IDE shows all tables and trigger
- Like https://github.com/hasura/graphql-engine
GraphQL Engine should be interpreter, but can code gen if needed.

Mutations go directly back into DB with zero code required

But need Validators somewhere. Can do that in Script.

Tech

CI https://github.com/ovh/cds

supports EVERYTHING
can run all desktops inside VSPhere.
Can run all in a single process. Check how they do that !!
Very nice compared to many others.

GPRC https://github.com/gunk/gunk-example-server

go --> GRPC --> All other languages.
Nice because you get easy support for all other languages.

GraphQL Introspection

https://github.com/graph-gophers/graphql-go
- GraphQL interpreted and well maintained
- Has subscriptions. https://github.com/graph-gophers/graphql-transport-ws
  - Ex: https://github.com/matiasanaya/go-graphql-subscription-example

GraphQL Middleware ( auth, tracing, etc etc) - https://github.com/rodrigobotelho/graphql-kit

Security Keys

https://github.com/mozilla/sops
- simple golang and allows them to be in Git.

CDC

Once in NATS, Benthos can be used to push into BLast or minio

B

Same as A, but:

No triggers and instead parsing WAL

How ?

Benthos does it.
Pump into NATS queues

C

Same as B, but:

We have a Read only Database

How ?

Use the WAL and Benthos for everything
Pump Data into BLast stores
GraphQL query on top of Blast with GraphQL IDE from the WAL like in A
Mutations go back to DB like in A.