wolkenkit is an open-source CQRS and event-sourcing framework based on Node.js, and it supports JavaScript and TypeScript.
Home Page: https://www.thenativeweb.io/wolkenkit/framework
License: GNU Affero General Public License v3.0
![dependabot-preview[bot] avatar](https://avatars.githubusercontent.com/in/2141?v=4 "dependabot-preview[bot]")
![dependabot[bot] avatar](https://avatars.githubusercontent.com/in/29110?v=4 "dependabot[bot]")
Currently we require the user to setup Docker. There are cases where this does not work without issues, e.g. when you try to do this on Windows and you have VirtualBox installed (you can't run VirtualBox and Hyper-V at the same time, and Hyper-V is required by Docker).
So we came up with the idea of having ready-made VMs for VirtualBox, VMware & co., that include Docker, Node.js and wolkenkit. This way people can fallback to these tools, but do not need to setup their VM manually. To be able to setup these ready-made VMs, we should write a Packer script, so we can automate the process easily.
wolkenkit user with wolkenkit passworden-US keyboard layout and timezone as PSTnvm and the latest Node.js LTSnpm install -g wolkenkitwolkenkit/wolkenkitCurrently, we handle wolkenkit primarily as a black box. It may be difficult for others to understand what is happening under the hood, and to understand the architecture. We should add more information on this in the documentation.
This was suggested by @schmuto, so we should inform him once this has been done. Additionally, we should also inform whotheheck and Sven in Slack.
As reported by @blumatixgs in #3 npm run start doesn't work after having run wolkenkit init on Windows 10.
@bluematixgs Is the following description correct, and can you please elaborate what fails means in the Actual result section?
wolkenkit init in an empty directorynpm run startnpm run start fails.I would extend the CLI Information Message by echoing "Docker Client missing - did you install the Docker Application on your Machine?" when he tries the run the "wolkenkit start" command, to tell the user that he is missing something.
I would also additionally create a "Prerequisites" category for https://github.com/thenativeweb/wolkenkit. (if there are more prerequisites required to work with wolkenkit)
Seems like mobile browser throttle client side JS which causes Websocket Connections to freeze while being inactive.
How does wolkenkit store aggregates? Inmemory/database.
Does it replay all the events from the database everytime the sever starts up or loads it from a snapshot?
Does it cache the events/aggregates on the server then what happens if we run our of memory? Does it load only the hot data?
Does this implement something like Prevalent System Architecture?
How much of a performance boost will this architecture have over a normal MVC/CRUD app without caching at all?
Is there someway we can expose a REST API/ GraphQL api using wolkenkit as the data provider?
So far, we marked MongoDB support as experimental in the wolkenkit-eventstore module. The reason for this is that MongoDB does not support transactions, which are needed to guarantee an all-or-nothing semantics when saving multiple events. Now MongoDB 4.0 introduced transaction support, but only for clusters. We should define whether we further want to support MongoDB 4.0, or if we get rid of it (since relational databases seem to be a better fit for event stores anyway than NoSQL databases).
Plans to make it open source? ๐
Using the where clause it is currently not possible to run a full-text search on a list, e.g. if you want to find all todos that start with the word Might. The currently available operators are focused on arrays (contains) or simple data types (greaterThan).
MongoDB, which we currently use as the default database for read models, has two possibilities to deal with situations like these:
Maybe it would be enough to add support for these operators to the translate functions inside the broker.
$text and $regex operators from MongoDB workYou might also want to see this StackOverflow question.
Would it be possible to publish the wolkenkit roadmap and/or feature release dates?
From time to time, people ask how wolkenkit compares to other BaaS platforms. We could add a section to the documentation where we contrast wolkenkit to some of its alternatives, similar to the way HashiCorp compares their tools to others (see the Consul documentation as an example).
We have a number of problems with the current HTTP/2 endpoint in wolkenkit:
local.wolkenkit.io must be considered compromised.After having discussed this we came to the conclusion that the only viable solution is to introduce an HTTP endpoint, which becomes the new default, and have an additional optional HTTP/2 endpoint. This issue describes the steps needed to implement this.
We might rate limit the API using a module such as express-brute (or something similar, this is just a suggestion, not a specifically chosen module).
This will probably need to be done inside of the tailwind module, maybe somewhere around here.
I set up Wolkenkit 1.2.0 on Windows 10 Professional as described in the documentation. When I run wolkenkit start there's an error:
$ wolkenkit start
Starting the application...
Verifying health on environment default...
โ events.js:183
throw er; // Unhandled 'error' event
^
Error: write EPIPE
at _errnoException (util.js:1024:11)
at WriteWrap.afterWrite [as oncomplete] (net.js:867:14)
The same error occurs when I'm trying to start a Wolkenkit template or when I'm using the wolkenkit health command. The Hyper-V machine is up and running and responds to pings.
wolkenkit initwolkenkit startTo help understand the use cases for wolkenkit, we should setup an additional page Not a developer (this may also be interesting for https://www.wolkenkit.io/). There we should list good and bad use cases, and reasons for the management how a company can benefit from using wolkenkit.
We have made up the following items (they still need to be discussed, and maybe shortened):
At present, wolkenkit only supports OpenID Connect for auth. There will be cases where this might not be feasible/relevant, e.g. existing authentication/authorization microservices, etc.
When starting a wolkenkit application, we run npm install internally. Right now this runs as anonymous, which means that you can only install publicly available modules. To support private modules, the user would need to login to npm first. We could ask the user to provide an npm token, which we then use internally to run npm install.
npm installIn your Wolkenkit-Brochure you wrote on page 48:
calculate the new number of likes and add this information when publishing the liked event:
// ... message.events.publish('liked', { likes: message.state.likes + 1 }); // ...
Does the Wolkenkit framework ensure this also works correctly when two users are liking a message "at the same time"?
E.g.
message.state.likes is not yet updated)message.state.likes is updated to 1message.state.likes is updated to 1I think this boils down to the question when the state is updated, and if two commands may be executed in parallel.
Udo
It might be interesting to have an option to encrypt some data, e.g. when storing passwords. This would have to be done in the event store as well as in the read model. We should think about a way of how to do this.
This issue came up in a discussion with @reneviering, so we should inform him once this has been discussed.
After setting up a new project with wolkenkit init on my Ubuntu 16.04 machine, wolkenkit start fails with the message
โ Failed to create container chat-broker. Error response from daemon: invalid IP address in add-host: ""
I'm running docker v17.05.0-ce.
Any help is much appreciated!
For some applications it would make sense to run logic on the server when special things happen, e.g. when a client disconnects. Unfortunately, this can't be implemented by sending a command from the client, since the user may just close the browser. Although there is an unbeforeunload event in the browser for things like that, this event does not work reliably. There may be more events like this.
To be able to handle system events and domain events in a similar way, we might introduce a new context for them, such as app, so that you get events similar to app.client.disconnected.
@colorizedmind was interested in this, so we should inform him once this has been implemented.
There are two commands in the CLI that have been marked as deprecated a while ago. For the next release of the CLI we could finally get rid off them. The affected commands are:
updatels-remoteRight now, the changelog contains a list of changes made in a specific version. This list is ordered by criteria such as added, updated, and removed, but this leads to a quite technical result. It would be nice to do this in a more feature-driven way. Apart from that, it would be great if we could link to individual changes in the changelog.
latest to the new conceptlatest)Please help me with clearing the database from old, erroneous data.
I'm guessing there are access credentials needed to 'clear' the Events
How about the denormalised data, thats stored in mongo?
What process should i follow to for example, remove old test data?
Since Docker runs on Raspberry Pi, it would be great if you could also run wolkenkit on Raspberry Pi. This would allow us to provide an image that you can download that contains Docker and all the required wolkenkit Docker images. This could be a highly interesting thing for tinkering and IoT solutions.
How could we integrate wolkenkit to a react based app? An example would be nice.
any help please
Currently, we don't have a single guide to describe how to setup authentication. Using the blog post that we have written for Auth0 as a foundation, it should be pretty simple to setup such a guide. Additionally, this check list on StackOverflow might be helpful.
@schmuto has indicated interest on this, so we should inform him once this has been done.
The current example code is for very basic use cases. It would be great if you could include a more real-world example, with more complex entities, etc.
Right now, we only have a single anonymous identity for all anonymous users. We should introduce an option to be able to distinguish between individual anonymous users, e.g. to allow shopping in an online store without the need to login first.
Currently, when the broker can not successfully verify a token, it silently fails and returns a 401. The client then tries to re-authenticate itself. If there is something broken with how the client gets the token, this will fail and end in an endless loop. To discover these problems, a log message in the server logs would be very helpful.
Since the broker uses tailwind under its hood, this would probably need to be fixed there. The correct place seems to be the call to the limes middleware. Maybe this also requires some changes to limes, so that you get better errors back.
The Docker images and containers used in wolkenkit are named after metaphors instead of what they actually are responsible for. This way there is always the need to explain what e.g. the core is responsible for (the write model). We came up with the idea to maybe change the names of those images and containers to be more explicit here.
Right now, files that have been uploaded to wolkenkit's file storage are either configured to be readable by anyone, only by authenticated users, or only by their owner. While this works well for files that are handled programmatically, it gets difficult to show protected images, since you can't send a token with an <img> tag. Hence it would be great if we could add one-time links for downloads, that can be consumed a single time.
There are multiple approaches on how to solve this problem. Creating one-time links is only one option, another one could be to introduce a special function to the client SDK, maybe even a React component or something like this. But handling it in the backend is the only option that works independent of a specific SDK (since you may come from another programming language, or even work without the SDK).
Statefull flow with 2 transitions that listen to the same event.
'use strict';
const identity = {
'communication.message.liked': event => event.aggregate.id,
};
const initialState = {
is: 'pristine',
};
const transitions = {
pristine: {
'communication.message.liked'(flow, event) {
console.log('eventId:', event.id);
}
},
magic: {
'communication.message.liked'(flow, event) {}
}
};
const reactions = {
pristine: {
magic(flow, event, {app}) {}
},
magic: {
magic(flow, event, {app}){}
}
};
module.exports = {identity, initialState, transitions, reactions};
console.log should be called once with the event.id
console.log get called 2 times for every liked event
When you remove the magic transistion console.log is colled once for every liked event.
getClassifiedFlows creates two instances of the same statefull flow, where both react to the same event.
Following the Tutorial in documentation initializing-the-application I get the error "Template not found" when executing "wolkenkit init".
Console:
wolkenkitdev@wolkenkitdev ~/jsdoc $ cd chat/
wolkenkitdev@wolkenkitdev ~/jsdoc/chat $ ls -l
total 0
wolkenkitdev@wolkenkitdev ~/jsdoc/chat $ wolkenkit init
โ Failed to initialize a new application.
Template not found.
This is running with a linux mint installation within a VirtualBox VM on Windows Host. Network is configured as NAT. browsing the web (and installation) worked without setting proxy configuration (so I did not touch them).
I am setting up wolkenkit and going through the tutorial.
I cannot start the demo application.
Don't know if it's reproducible but I'm going through the tutorial:
wolkenkit initwolkenkit startThe application stops at Building Docker images....
I got docker images but no containers.
The demo application starts without issues.
This is the output if I run wolkenkit start --verbose
Starting the application...
Verifying health on environment default...
Application host local.wolkenkit.io resolves to 127.0.0.1.
Docker server resolves to 127.0.0.1 and ::1.
Verifying application status...
Verifying that ports are available...
Setting up network...
Building Docker images...
โ Failed to start the application.
unable to prepare context: path "/tmp/p1RZVL" not found
ExecutableFailed: unable to prepare context: path "/tmp/p1RZVL" not found
at new CustomError (/home/cwalther/.nvm/versions/node/v10.12.0/lib/node_modules/wolkenkit/node_modules/defekt/dist/defekt.js:44:116)
at /home/cwalther/.nvm/versions/node/v10.12.0/lib/node_modules/wolkenkit/dist/shell/exec.js:29:18
at ChildProcess.exithandler (child_process.js:296:5)
at ChildProcess.emit (events.js:182:13)
at maybeClose (internal/child_process.js:962:16)
at Socket.stream.socket.on (internal/child_process.js:381:11)
at Socket.emit (events.js:182:13)
at Pipe._handle.close (net.js:606:12)
I am using:
This will make it easier to identify which packages (officially) belong to wolkenkit.
@wolkenkitI'm using Docker for windows like the Docker inc. suggested url
Is it possible to add an installation guide like the "Installing using Docker Machine" but for "Installling using Docker with Hyper-V"?
When I try to execute the restart command and wolkenkit app is not running I get an error message that my app is not running and it is not starting. From similar commands of other CLIs I would expect it to start anyway.
wolkenkitdev@wolkenkitdev ~/jsdoc/chat $ wolkenkit restart
โ chat is not running.
In the past we had an issue due to a function call that was done in a wrong way. Instead of
app.events.observe({
where: { name: 'mounted' }
})โฆwe had:
app.events.observe({
name: 'mounted'
})โฆEssentially, the where keyword was missing. Maybe we should improve error checking of any public functions that are user-facing, so that things like that result in a meaningful error message.
At https://docs.wolkenkit.io/ we have the documentation for wolkenkit itself, but the APIs of the wolkenkit applications do not yet have any documentation. We could use the landing page of the broker to provide some automatically generated documentation based on the metadata of the model. This could help us to get closer to a self-documenting API.
/v1/configuration.json route to contain these informationHi there!
I first heard about Wolkenkit on a podcast and again on Hacker News a few days ago, so I decided to read the brochure. (I've previously enjoyed Martin Kleppmann's talks around Event Sourcing / CQRS.)
I'm a little confused by a few details in the brochure at this point:
When the 'like message' command is defined, it is used to increment the number of likes and publishes that new like count in the event. Wouldn't it be cleaner if the command did only validation and the event did the actual state update (including calculating the new like count)?
I think I'm not 100% sure where to draw the boundary between command and event and why.
And about the authorization code: Why do both commands and events need to have authentication rules defined? I would have expected that only an authorized command can publish a matching event.
Very cool framework!
Rainer
As reported by @blumatixgs in #3 there is an issue with running wolkenkit applications after a machine reboot on Windows 10.
@blumatixgs Could you please have a look at this issue, and elaborate on what exactly not reachable means? What exactly happens when you try to reach the application?
The wolkenkit application is running and reachable.
The wolkenkit application is running, but not reachable.
Hi guys, is there any TypeScript types created for this lib?
Thanks in advance.
According to the docs, environment variables are stored in the package.json file. I believe that it might be better to move these into a separate .env file. In 12-factor philosophy, that will be helpful in keeping the environment variables out of the code (and repo) and belong to the environment.
I would extend the CLI Information Message by telling the user "Docker Client missing - did you install the Docker Application on your Machine?" when he tries the run the command "wolkenkit start", to tell him that he is missing something.
I would also additionally create a "Prerequisites" category for https://github.com/thenativeweb/wolkenkit. (if there are more prerequisites required to work with wolkenkit)
Currently, from within a flow or from within the client you can use the asUser property when sending a command, to impersonate, i.e. to send the command under a different user. Technically, we only change the sub claim, we don't issue a real token. This works if you only want to override the user's identity, but sometimes you also need to provide additional claims. So it would be better to issue a real token, with a new sub claim as well as custom claims.
asUser functions to optionally handle custom claims
can-impersonate claim
This feature request was made by @schmuto. So we should notify him once this has been done.
I've added a new command to the chat template app called 'comment'.
It's supposed to add comments to chat messages. I've setup all the command, event and model things to reflect this.
However the UI complains in the console that my comment function does not exist.
Here it is, i've added it to the commands object:
comment (message, command, mark) {
message.events.publish('commented', {
comments: message.state.comments.push(command.data.comment)
});
mark.asDone();
}
The Error (if that helps):
index.js:28 Uncaught TypeError: chat.communication.message(...).comment is not a function
at HTMLUListElement.view.messages.addEventListener.event
Currently the developer of a wolkenkit application has to download the certificate for the identity provider manually, and store it somewhere. This could be automated using JWKS.
package.jsonThis was suggested by @AlexZeitler, so we should inform him once this has been implemented.
Right now there is no option in the wolkenkit CLI to backup and restore an application's data, such as the events from the event store and the files from depot. To allow backing up and restoring, and maybe also migrating data from one system to another, we should introduce two new commands to the CLI, export and import:
$ wolkenkit export --env default --to <directory> --from-event-store true
$ wolkenkit import --env default from <directory>default as the default environmenttrue as the default value for from-event-store<directory>/event-store
events-0000000000000001.json (and so on)default as the default environment
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft
Open source projects and samples from Microsoft.
Google
Google โค๏ธ Open Source for everyone.
Alibaba
Alibaba Open Source for everyone
D3
Data-Driven Documents codes.
Tencent
China tencent open source team.