prooph/micro

Functional prooph for microservices

prooph-micro

[Experimental] Functional prooph for microservices

Installation

composer install

Example usage

php examples/register_and_change_username.php

Run example snapshotter

php examples/user_snapshotter.php

prooph Microservices in Action

prooph/micro-do

Support

• Ask questions on Stack Overflow tagged with #prooph.
• File issues at https://github.com/prooph/micro/issues.
• Say hello in the prooph gitter chat.

Contribute

Please feel free to fork and extend existing or add new plugins and send a pull request with your changes! To establish a consistent code quality, please provide unit tests for all your changes and may adapt the documentation.

License

Released under the New BSD License.

micro's Issues

Add message validator

Let's check a message upfront and return early, when it's not confirming to the requirements

Make use of pthreads

Is it worth to threat?

Consider nginx unit

see http://unit.nginx.org/docs-nginx-unit.html

thanks for the hint @max-voloshin

Integerate php-pm

Maybe we should integrate https://github.com/php-pm/php-pm by default?

add micro command for composer

Requirements:

• able to run composer update for a single microservice
• able to run composer install for a single microservice
• able to run composer update for all microservices
• able to run composer install for all microservices

Example docker-command:

docker run --rm -it --volume $(dirname $SSH_AUTH_SOCK):$(dirname $SSH_AUTH_SOCK) -e SSH_AUTH_SOCK=$SSH_AUTH_SOCK --volume $(pwd):/app --volume $(pwd)/../../packages:/packages prooph/composer:7.1 install

As you can see, this docker command is quite long and hard to remember, therefore we want to provide shortcuts for micro.

Things to note:

Each php-service can use a different image and PHP version (f.e. 5.6 on service1 and 7.1 on service2). Therefore the correct prooph/composer images with the same php version has to be chosen correctly by this command.

Syntax suggestion

./bin/micro micro:composer:update service1
./bin/micro micro:composer:update --all
./bin/micro micro:composer:install service2
./bin/micro micro:composer:install --all

Move cli + folder structure to a skeleton repo

Reason: You clone the skeleton ones and modify it to your needs, but you may wish to update micro as a library like any other composer dependency.

Consider Serverless Event Gateway

https://github.com/serverless/event-gateway

Looks like micro and event gateway would play very nice together.

Input and output streams can differ

I often use the concept that one aggregate starts the lifecycle of another aggregate.

We have such an example in prooph-do: https://github.com/prooph/proophessor-do/blob/master/src/Model/Todo/Handler/PostTodoHandler.php#L57

user aggregate handles PostTodo command and a new todo aggregate is born.
How can we handle this without a dedicated command handler, configure "read from stream" and "write to stream" separately?

TypeError on example

php examples/register_and_change_username.php

Gives an TypeError. Not sure if that is intentional cause running an InvalidCommand

TypeError
"Return value of Prooph\\Common\\Messaging\\DomainMessage::messageName() must be of the type string, null returned"

Add akka-like messaging

from the chat:

an aggregate talks to the process manager (or another aggregate) directly, no magic wiring in the background. that's a huge difference, and I like the explicitness - would like to have something like this in micro - thoughts?

Remove AmqpReducer

With #23 publishing was removed as step of the micro pipeline, so it is left to the user how to publish. Hence, we should remove the AmqReducer to not couple micro to a specific message queue

Version vs. event no problem with snapshots

We have a larger problem maybe with the new snapshot implementation in general. See this line: https://github.com/prooph/micro/blob/master/src/SnapshotReadModel.php#L79

This is wrong. You cannot query the event position using the version of the aggregate because they are not the same.
We cannot use the aggregate version to query missing events from the event stream.

See here: https://github.com/prooph/micro/blob/master/src/Kernel.php#L81
At the moment information from snapshot store is not used but how can we change that?
We would need to store the event position along with the snapshot but we don't have this information available in the message.

Any ideas @prolic

Optimize for one stream per aggregate

In this case we don't need all this metadata matching and can query the stream position directly.

[Question] Should we remove state and snapshots?

While writing the Event Sourcing part of our walk-through guide I noticed that it is difficult to explain why we pass state to an aggregate function. A few questions came to mind: Where does the state come from (of course I know the answer but it requires a lot of explanation)? Why don't we work with the history instead?

I did a short research and found this excellent blog post by @mathiasverraes. I read the post long time ago but my focus was on object oriented ES and not the fp way.

Mathias uses f(history, command) => events to describe an aggregate function.

Our definition looks a bit different atm: f(state, command) => newState // events

It allows us to work with state snapshots and the aggregate function can inspect state to verify that command can be processed. But the aggregate function could also inspect history instead.

If we pass history and only return events our aggregate functions become even more simple. We no longer need the apply method as this would be only part of a projection. The aggregate can iterate history and filter events with required information. As long as the history of a single aggregate is not too big it would work quite well.

Add snapshot support

Remove upstream definition from nginx site config

The upstream definition is only needed to add multiple server for one upstream. But this should be handled by an load balancer who resolves the Docker service name. This is done by Docker Swarm or an AWS Elastic Load Balancer. Does we really need it? But I have to check if the resolver definition is needed then, this would be a problem.

Update Documentation

Change command handlers

function (callable $stateResolver, Command $command): Events[]

Pros:

• more explicit
• no loading of state / history events, when not needed

refers to #34

Get aggregate state version by counting recorded events

ATM we require the developer to set new version: https://github.com/prooph/micro/blob/master/examples/Model/User.php#L32
I'd like to get rid of it if possible.

add amqp publisher tests

micro:setup:php-service cli command -> start command not working

When adding a run command:

Enter the start command (f.e. "php run.php"):  []:
 > php run.php

This is not reflected in the docker-compose.yml

Add pdo env to php images if they are depended

Scenario

1. User creates a postgres/mysql database (./bin/micro micro:setup:postgres)
2. User creates a php-service and selects this postgres database as dependend (./bin/micro micro:setup:php-service)

Feature Request

All MySQL and PostgreSQL databases created by micro cli commands, have a label prooph-pdo:true in the docker-compose.yml.

When the php-service is created, it should check for the prooph-pdo label and add the following environment variables to the php-image:

PROOPH_PDO_DSN=pgsql:host=postgres port=5432 dbname=micro_do
PROOPH_PDO_USER=postgres
PROOPH_PDO_PWD=

with the values of host name, port number, dbname, user & password taken from the docker-compose.yml file

Goal

In the PHP service you can create a database connection very easy with the given environment variables, so no additonal config is required for this service.