* update dapr runtime to use v1.12.0-rc.1 Signed-off-by: Cassandra Coyle <cassie@diagrid.io> * update to latest rc for runtime and cli Signed-off-by: Cassandra Coyle <cassie@diagrid.io> * try sleeping Signed-off-by: Cassandra Coyle <cassie@diagrid.io> * pin redis version Signed-off-by: Cassandra Coyle <cassie@diagrid.io> * rm sleep Signed-off-by: Cassandra Coyle <cassie@diagrid.io> * rm deprecated unsubscribe call Signed-off-by: Cassandra Coyle <cassie@diagrid.io> * Update configuration/go/sdk/order-processor/app.go Co-authored-by: Josh van Leeuwen <me@joshvanl.dev> Signed-off-by: Cassie Coyle <cassie.i.coyle@gmail.com> * Statemanagement multiapp run Signed-off-by: rabollin <rabollin@microsoft.com> * message sequence Signed-off-by: rabollin <rabollin@microsoft.com> * message sequence Signed-off-by: rabollin <rabollin@microsoft.com> * Multi-run support for workflow Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Removing component folder from run file to match readme Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * aligning state store names on `statestore` Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Adding multi-app run for python and C# Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * pr comments Signed-off-by: rabollin <rabollin@microsoft.com> * pr comments Signed-off-by: rabollin <rabollin@microsoft.com> * sleep value increase Signed-off-by: rabollin <rabollin@microsoft.com> * sleep seconds increase Signed-off-by: rabollin <rabollin@microsoft.com> * Result correction Signed-off-by: rabollin <rabollin@microsoft.com> * Result correction Signed-off-by: rabollin <rabollin@microsoft.com> * java sdk dapr run output format fixing Signed-off-by: rabollin <rabollin@microsoft.com> * Syncing Release 1.12 with master and multi-run changes (#936) * Add holopin.yml config (#923) Signed-off-by: Marc Duiker <marcduiker@users.noreply.github.com> * Updating Readme's to default to `dapr run -f .` (#874) * Updating Invoke readme in C# to favor dapr run -f Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MacBook-Pro-2.local> * Adding back single app run steps and tests, but to the end. Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Update service_invocation/csharp/http/README.md Co-authored-by: Mark Fussell <markfussell@gmail.com> Signed-off-by: Paul Yuknewicz <paulyuk@microsoft.com> * Updated and hardened test Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Invoke * and partial PubSub with multi-run files & readmes & tests Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Multi run done for Pubsub C#, Python and Node (http, sdk) Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MacBook-Pro-2.local> * Added Pubsub support for Go, Java Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MacBook-Pro-2.local> * updating readmes with correct appid Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MacBook-Pro-2.local> * Updating tests for PubSub and Service invoke to Paas with multiapp run Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * More fine tuning of pubsub tests (timeouts, step ends) Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Fixing Python FastAPI and C# Pubsub tests Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> --------- Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MacBook-Pro-2.local> Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> Signed-off-by: Paul Yuknewicz <paulyuk@microsoft.com> Co-authored-by: Paul Yuknewicz <paulyuk@Pauls-MacBook-Pro-2.local> Co-authored-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> Co-authored-by: Mark Fussell <markfussell@gmail.com> * Patching tests to pass (C# timing, Python uvicorn) (#933) Signed-off-by: Paul Yuknewicz <paulyuk@microsoft.com> --------- Signed-off-by: Marc Duiker <marcduiker@users.noreply.github.com> Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MacBook-Pro-2.local> Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> Signed-off-by: Paul Yuknewicz <paulyuk@microsoft.com> Co-authored-by: Marc Duiker <marcduiker@users.noreply.github.com> Co-authored-by: Paul Yuknewicz <paulyuk@Pauls-MacBook-Pro-2.local> Co-authored-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> Co-authored-by: Mark Fussell <markfussell@gmail.com> * Fixing MMD tests with tweaks Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * setting version for kafka (#939) * fixing version for kafka Signed-off-by: Pravin Pushkar <ppushkar@microsoft.com> * validate data with some initial values instead of larger Signed-off-by: Pravin Pushkar <ppushkar@microsoft.com> --------- Signed-off-by: Pravin Pushkar <ppushkar@microsoft.com> * Updating test env to Dapr runtime 1.12 RC4 and CLI RC2 Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Update README.md Fixing typo. should be singular Signed-off-by: Paul Yuknewicz <paulyuk@microsoft.com> * Update README.md Signed-off-by: Paul Yuknewicz <paulyuk@microsoft.com> * add java workflow quickstart Signed-off-by: Sky Ao <aoxiaojian@gmail.com> * Updating runtime for tests to RC5 Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Fixes port 6003 conflict between Go Pub-sub and Go bindings. Timing issue Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * add hello-kubernetes multi-app run k8s template Signed-off-by: Mukundan Sundararajan <65565396+mukundansundar@users.noreply.github.com> * address review comments. Signed-off-by: Mukundan Sundararajan <65565396+mukundansundar@users.noreply.github.com> * update cli to rc.4 (#943) * update cli to rc.3 Signed-off-by: Cassandra Coyle <cassie@diagrid.io> * update cli to rc.4 Signed-off-by: Cassandra Coyle <cassie@diagrid.io> --------- Signed-off-by: Cassandra Coyle <cassie@diagrid.io> Signed-off-by: Cassie Coyle <cassie.i.coyle@gmail.com> * Update CLI to rc4 Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Updating .NET SDK to 1.12.* Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Dapr Golang SDK update to v1.9.0-rc-1 Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Javascript SDK to 3.1.2 Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Release 1.12 (#945) * Dapr Golang SDK update to v1.9.0-rc-1 Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Javascript SDK to 3.1.2 Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> --------- Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> Co-authored-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Java SDK update to <version>1.10.0-rc-1</version> Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Updated Java SDK to RC, and donated multi-app-run support too :) Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> * Updating tests to use dapr/dapr 1.12.0-rc.6 Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MacBook-Pro-2.local> * sdk version changes Signed-off-by: rabollin <rabollin@microsoft.com> * CLI and Runtime version changes Signed-off-by: rabollin <rabollin@microsoft.com> * go.sum refresh Signed-off-by: rabollin <rabollin@microsoft.com> * Updating with js-sdk 3.2.0 Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> --------- Signed-off-by: Cassandra Coyle <cassie@diagrid.io> Signed-off-by: Cassie Coyle <cassie.i.coyle@gmail.com> Signed-off-by: rabollin <rabollin@microsoft.com> Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> Signed-off-by: Marc Duiker <marcduiker@users.noreply.github.com> Signed-off-by: Paul Yuknewicz <paulyuk@Pauls-MacBook-Pro-2.local> Signed-off-by: Paul Yuknewicz <paulyuk@microsoft.com> Signed-off-by: Pravin Pushkar <ppushkar@microsoft.com> Signed-off-by: Sky Ao <aoxiaojian@gmail.com> Signed-off-by: Mukundan Sundararajan <65565396+mukundansundar@users.noreply.github.com> Co-authored-by: Cassandra Coyle <cassie@diagrid.io> Co-authored-by: Cassie Coyle <cassie.i.coyle@gmail.com> Co-authored-by: Josh van Leeuwen <me@joshvanl.dev> Co-authored-by: rabollin <rabollin@microsoft.com> Co-authored-by: Paul Yuknewicz <paulyuk@Pauls-MBP-2.lan> Co-authored-by: Marc Duiker <marcduiker@users.noreply.github.com> Co-authored-by: Paul Yuknewicz <paulyuk@Pauls-MacBook-Pro-2.local> Co-authored-by: Mark Fussell <markfussell@gmail.com> Co-authored-by: Pravin Pushkar <ppushkar@microsoft.com> Co-authored-by: Sky Ao <aoxiaojian@gmail.com> Co-authored-by: Mukundan Sundararajan <65565396+mukundansundar@users.noreply.github.com> |
||
---|---|---|
.. | ||
csharp | ||
deploy | ||
go | ||
img | ||
node | ||
python | ||
react-calculator | ||
README.md | ||
makefile | ||
operands.json | ||
persist.json |
README.md
Distributed calculator
This quickstart shows method invocation and state persistent capabilities of Dapr through a distributed calculator where each operation is powered by a different service written in a different language/framework:
- Addition: Go mux application
- Multiplication: Python flask application
- Division: Node Express application
- Subtraction: .NET Core application
The front-end application consists of a server and a client written in React.
Kudos to ahfarmer
whose React calculator
The following architecture diagram illustrates the components that make up this quickstart:
Prerequisites for running the quickstart
Clone the quickstarts repository
git clone [-b <dapr_version_tag>] https://github.com/dapr/quickstarts.git
Note: See https://github.com/dapr/quickstarts#supported-dapr-runtime-version for supported tags. Use
git clone https://github.com/dapr/quickstarts.git
when using the edge version of dapr runtime.
- Run locally
- Run in Kubernetes environment
- Dapr-enabled Kubernetes cluster. Follow these instructions to set this up.
Running the quickstart locally
These instructions start the four calculator operator apps (add, subtract, multiply and divide) along with the dapr sidecar locally and then run the front end app which persists the state in a local redis state store.
- Add App - Open a terminal window and navigate to the go directory and follow the steps below:
- Install the gorilla/mux package: Run:
go get -u github.com/gorilla/mux
- Run dapr using the command:
dapr run --app-id addapp --app-port 6000 --dapr-http-port 3503 go run app.go
- Subtract App - Open a terminal window and navigate to the csharp directory and follow the steps below:
-
Set environment variable to use non-default app port 7001
#Linux/Mac OS: export ASPNETCORE_URLS="http://localhost:7001"
#Windows: set ASPNETCORE_URLS=http://localhost:7001
- Build the app. Run:
dotnet build
- Navigate to ./bin/Debug/netcoreapp7.0 and start Dapr using command:
dapr run --app-id subtractapp --app-port 7001 --dapr-http-port 3504 dotnet Subtract.dll
- Divide App - Open a terminal window and navigate to the node directory and follow the steps below:
- Install dependencies by running the command:
npm install
- Start Dapr using the command below
dapr run --app-id divideapp --app-port 4000 --dapr-http-port 3502 node app.js
- Multiply App - Open a terminal window and navigate to the python directory and follow the steps below:
- Install required packages
pip3 install wheel python-dotenv flask_cors flask
- Set environment variable to use non-default app port 5000
#Linux/Mac OS: export FLASK_RUN_PORT=5001 #Windows: set FLASK_RUN_PORT=5001
- Start dapr using the command:
dapr run --app-id multiplyapp --app-port 5001 --dapr-http-port 3501 flask run
- Frontend Calculator app - Open a terminal window and navigate to the react-calculator directory and follow the steps below:
- Install the required modules
npm install npm run buildclient
- Start Dapr using command below:
dapr run --app-id frontendapp --app-port 8080 --dapr-http-port 3507 node server.js
-
Open a browser window and go to http://localhost:8080/. From here, you can enter the different operations.
-
Open your browser's console window (using F12 key) to see the logs produced as you use the calculator. Note that each time you click a button, you see logs that indicate state persistence and the different apps that are contacted to perform the operation.
- Optional: Curl Validate
-
To make sure all the apps are working, you can run the following curl commands which will test all the operations:
curl -s http://localhost:8080/calculate/add -H Content-Type:application/json --data @operands.json
curl -s http://localhost:8080/calculate/subtract -H Content-Type:application/json --data @operands.json
curl -s http://localhost:8080/calculate/divide -H Content-Type:application/json --data @operands.json
curl -s http://localhost:8080/calculate/multiply -H Content-Type:application/json --data @operands.json
curl -s http://localhost:8080/persist -H Content-Type:application/json --data @persist.json
curl -s http://localhost:8080/state
- You should get the following output:
86 18 1.5294117647058822 1768 {"operation":null,"total":"54","next":null}
- Cleanup
-
Cleanup microservices
dapr stop --app-id addapp
dapr stop --app-id subtractapp
dapr stop --app-id divideapp
dapr stop --app-id multiplyapp
dapr stop --app-id frontendapp
- Uninstall node modules by navigating to the node directory and run:
npm uninstall
Running the quickstart in a Kubernetes environment
- Navigate to the deploy directory in this quickstart directory:
cd deploy
Note:
appconfig.yaml
is not used directly for this quickstart but is present for the observability quickstart. - Follow these instructions to create and configure a Redis store
- Deploy all of your resources:
kubectl apply -f .
Note: If you had previously installed, dapr using the
dapr init -k --dev
command, theappconfig
previously installed, would have changed with the above command to point to a different zipkin collector. To point it back to the zipkin collector installed duringdapr init -k --dev
, replace theendpointAddress
inappconfig.yaml
with the following:endpointAddress: "http://dapr-dev-zipkin.default.svc.cluster.local:9411/api/v2/spans"
and run the following command:
kubectl apply -f appconfig.yaml
Note: Services could also be deployed one-by-one by specifying the .yaml file:
kubectl apply -f go-adder.yaml
.
Each of the services will spin up a pod with two containers: one for your service and one for the Dapr sidecar. It will also configure a service for each sidecar and an external IP for the front-end, which allows us to connect to it externally.
- Kubernetes deployments are asyncronous. This means you'll need to wait for the deployment to complete before moving on to the next steps. You can do so with the following commands:
kubectl rollout status deploy/addapp
kubectl rollout status deploy/subtractapp
kubectl rollout status deploy/divideapp
kubectl rollout status deploy/multiplyapp
kubectl rollout status deploy/calculator-front-end
You can view the status of the running pods with:
kubectl get pods
When everything is running properly, you'll see output like this:
NAME READY STATUS RESTARTS AGE
addapp-5ff9586df6-5bpll 2/2 Running 0 16s
calculator-front-end-56dc959b58-bb8vw 2/2 Running 0 16s
divideapp-c64f744d6-wljcc 2/2 Running 0 16s
multiplyapp-6989454d77-tkd6c 2/2 Running 0 16s
subtractapp-869b74f676-9mw94 2/2 Running 0 16s
- Next, setup access to your service
There are several different ways to access a Kubernetes service depending on which platform you are using. Port forwarding is one consistent way to access a service, whether it is hosted locally or on a cloud Kubernetes provider like AKS.
kubectl port-forward service/calculator-front-end 8000:80
This will make your service available on http://localhost:8000. Navigate to this address with your browser and voilà! You have a working distributed calculator!
Optional: If you are using a public cloud provider, you can substitue your EXTERNAL-IP address instead of port forwarding. You can find it with:
kubectl get svc
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
dapr-api ClusterIP 10.103.71.22 <none> 80/TCP 135m
dapr-placement ClusterIP 10.103.53.127 <none> 80/TCP 135m
dapr-sidecar-injector ClusterIP 10.104.220.35 <none> 443/TCP 135m
addapp-dapr ClusterIP 10.0.1.170 <none> 80/TCP,50001/TCP 2m
calculator-front-end LoadBalancer 10.0.155.131 40.80.152.125 80:32633/TCP 3m
calculator-front-end-dapr ClusterIP 10.0.230.219 <none> 80/TCP,50001/TCP 3m
divideapp-dapr ClusterIP 10.0.240.3 <none> 80/TCP,50001/TCP 1m
kubernetes ClusterIP 10.0.0.1 <none> 443/TCP 33d
multiplyapp-dapr ClusterIP 10.0.217.211 <none> 80/TCP,50001/TCP 1m
subtractapp-dapr ClusterIP 10.0.146.253 <none> 80/TCP,50001/TCP 2m
Each service ending in "-dapr" represents your services respective sidecars, while the calculator-front-end
service represents the external load balancer for the React calculator front-end.
- Open your browser's console window (using F12 key) to see the logs produced as you use the calculator. Note that each time you click a button, you see logs that indicate state persistence:
Persisting State:
{total: "21", next: "2", operation: "x"}
total
, next
, and operation
reflect the three pieces of state a calculator needs to operate. The app persists these to a Redis store (see Simplified State Management section below). By persisting these, you can refresh the page or take down the front-end pod and still jump right back where you were. Try it! Enter something into the calculator and refresh the page. The calculator should have retained the state, and the console should read:
Rehydrating State:
{total: "21", next: "2", operation: "x"}
Also note that each time you enter a full equation (e.g. "126 ÷ 3 =") the logs indicate that a call is made to the service:
Calling divide service
The client code calls to an Express server, which routes the calls through Dapr to the back-end services. In this case the divide endpoint is called on the nodejs application.
- Optional: If your environment doesn't have easy access to a browser, or you just like using curl
Then you can use the following curl commands to make sure each one of the microservies is working:
curl -s http://localhost:8000/calculate/add -H Content-Type:application/json --data @operands.json
curl -s http://localhost:8000/calculate/subtract -H Content-Type:application/json --data @operands.json
curl -s http://localhost:8000/calculate/divide -H Content-Type:application/json --data @operands.json
curl -s http://localhost:8000/calculate/multiply -H Content-Type:application/json --data @operands.json
curl -s http://localhost:8000/persist -H Content-Type:application/json --data @persist.json
curl -s http://localhost:8000/state
You should get the following output:
86
18
1.5294117647058822
1768
{"operation":null,"total":"54","next":null}
Cleanup
Kubernetes environment cleanup
- Once you're done, you can spin down your Kubernetes resources by navigating to the
./deploy
directory and running:
kubectl delete -f .
This will spin down each resource defined by the .yaml files in the deploy
directory, including the state component.
The Role of Dapr
This quickstart demonstrates how to use Dapr as a programming model for simplifying the development of distributed systems. In this quickstart, Dapr is enabling polyglot programming, service discovery and simplified state management.
Polyglot programming
Each service in this quickstart is written in a different programming language, but they're used together in the same larger application. Dapr itself is language agnostic - none of the services have to include any dependency in order to work with Dapr. This empowers developers to build each service however they want, using the best language for the job or for a particular dev team.
Service invocation
When the front-end server calls the respective operation services (see server.js
code below), it doesn't need to know what IP address they live at or how they were built. Instead it calls their local dapr side-car by name, which knows how to invoke the method on the service, taking advantage of the platform’s service discovery mechanism, in this case Kubernetes DNS resolution.
The code below shows calls to the "add" and "subtract" services via the Dapr URLs:
const daprUrl = `http://localhost:${daprPort}/v1.0/invoke`;
app.post('/calculate/add', async (req, res) => {
const appResponse = await axios.post(`${daprUrl}/addapp/method/add`, req.body);
return res.send(`${appResponse.data}`);
});
app.post('/calculate/subtract', async (req, res) => {
const appResponse = await axios.post(`${daprUrl}/subtractapp/method/subtract`, req.body);
return res.send(`${appResponse.data}`);
});
...
Microservice applications are dynamic with scaling, updates and failures causing services to change their network endpoints. Dapr enables you to call service endpoints with a consistent URL syntax, utilizing the hosting platform’s service discovery capabilities to resolve the endpoint location.
Learn more about Dapr service invocation.
Simplified state management
Dapr sidecars provide state management. In this quickstart, the calculator's state is persisted each time a new button is clicked. This means a user can refresh the page, close the page or even take down the calculator-front-end
pod, and still retain the same state when they next open it. Dapr adds a layer of indirection so that the app doesn't need to know where it's persisting state. It doesn't have to keep track of keys, handle retry logic or worry about state provider specific configuration. All it has to do is GET or POST against its Dapr sidecar's state endpoint: http://localhost:3500/v1.0/state/${stateStoreName}
.
Take a look at server.js
in the react-calculator
directory. Note that it exposes two state endpoints for the React client to get and set state: the GET /state
endpoint and the POST /persist
endpoint. Both forward client calls to the Dapr state endpoint:
const stateUrl = `http://localhost:${daprPort}/v1.0/state/${stateStoreName}`;
Our client persists state by simply POSTing JSON key-value pairs (see react-calculator/client/src/component/App.js
):
const state = [{
key: "calculatorState",
value
}];
fetch("/persist", {
method: "POST",
body: JSON.stringify(state),
headers: {
"Content-Type": "application/json"
}
});
[Optional Steps] VS Code Debugging
If you are using Visual Studio Code, you can debug this application using the preconfigured launch.json and task.json files in the .vscode folder. The .vscode folder has already been modified in the project to allow users to launch a compound configuration called "Full Dapr App" which will run all applications and allow you to debug in VS Code.
For more information on how to configure the files visit How-To: Debug multiple Dapr applications
Note: You still need to edit your environment variables for Flask and ASPNETCORE_URLS Note: Dapr offers a preview Dapr Visual Studio Code extension for local development which enables users a variety of features related to better managing their Dapr applications and debugging of your Dapr applications for all supported Dapr languages which are .NET, Go, PHP, Python and Java.
Next Steps
- Explore additional quickstarts.