Azure Container Apps
This article goes over the deployment process for On-Premise solution of DecisionRules using Azure Container Apps.
Was this helpful?
Version 1.19.5 and Older
This article goes over the deployment process for On-Premise solution of DecisionRules using Azure Container Apps.
Was this helpful?
This tutorial uses Private Endpoints to communicate with the database, the cache and our key vault where we will keep secrets like certificates, license keys and other. If your deployment has no need for such security measures you can provision your resources as public, lowering the difficulty and making the deployment itself faster.
Following steps might differ depending on your level of security and sofistication of your existing Azure Cloud environment.
To follow this article successfully you will need the following things:
An Azure account with a Subscription
Permissions to create and manage the resources we're going to be using
Permissions to manage Microsoft Entra ID, groups and RBAC (for SSO and Key Vault setup)
It is also recommended to have these prepared:
Resource groups that fit with your conventions
Below are the steps our deployment will follow.
Navigate to Virtual networks -> Create virtual network
Fill in your Resource group, name and Region.
When creating the vnet consider creating a dedicated subnet for the Container Apps Environment - althogh this can be done later when creating the environment itself.
Depending on your provisioned resources these values might look a lot different.
Leave the rest of the options default (unless specified otherwise). You may want to provide tags for easier management. Create the Virtual Network.
Storing your connection strings, license key and SAML Certificate in a key vault is optional but highly recommended.
Navigate to Key vaults -> Create key vault. Fill in the required fields, for Recovery options you can choose whatever suits your use-case best.
In Access configuration the recommended approach is to use Azure RBAC. If, however, you don't have permissions to manage the RBAC policies of your company's cloud, you can go with Vault access policy.
In networking I choose to disable public access and create a Private Endpoint.
Review the key vault settings and create.
In the newly created key vault, navigate to Objects / Secrets. Click Generate/Import and name your secrets in recognizable ways and provide the values.
I recommend creating 4 secrets:
The license key for DecisionRules
The certificate for SAML-based authentication
The connection string for Cosmos DB
The connection string for Azure Cache for Redis
If you have a key rotation schedule you can implement it while creating the secrets by setting activation/expiration dates.
Navigate to Container Apps and hit Create. You will be presented with a familiar settings screen. First we have to create the Container Apps Environment.
In the Basics section of the Container Apps Creation screen under Container Apps Environment hit Create new.
In the Basics part of the Container Apps Environment screen provide the environment with a name. Select Workload Profiles for the Environment type and choose your preferred settings for Zone redundancy.
Workload profiles tab will stay default and for Monitoring you can choose whatever your use-case needs. Keep in mind not having logs will be a hinderance if you ever need to debug the infrastrucure.
In the Networking tab select Yes to use your own virtual network.
Assign your virtual network and either provide your previosly created subnet or click Create new to provision it now. Virtual IP settings should be kept External. If you decide to go with Internal for security reasons the accessibilty of the DecisionRules application will be extremely limited. Also keep in mind you will atleast have to create an outbound Firewall rule for the server containers calls to our license server.
Create the environment.
Deselect Use quickstart image and give your container a name. For Image source select Docker Hub or other registries and for Image type select Public.
The Image and tag are going to be the following:
Client container: decisionrules/client:<YOUR_PREFFERED_VERSION>
Server container: decisionrules/server:<YOUR_PREFFERED_VERSION>
BI container: decisionrules/business-intelligence:<YOUR_PREFFERED_VERSION>
Container resource allocation:
Client container: min. 0.5 CPU cores, 1 Gi memory
Server container: min. 1 CPU cores, 2 Gi memory
BI container: min. 1 CPU cores, 2 Gi memory
Keep the Bindings tab default and continue to the Ingress part of our setup.
Ingress has to be enabled for all of the containers. Traffic should be accepted from anywhere using the HTTP type. Ignore Client certificate mode.
Target ports for containers:
Client container: 80
Server container: 8080
BI container: 8082
Create all of your containers using the settings outlined unless your use-case demands otherwise.
At this point your Container Apps should look something like this:
Navigate to the container's Settings / Identity. Turn Status to On. This has to be done for containers that need access to our key vault.
When that's finished applying, go to your key vault's Access control (IAM) and add a role assignment. Choose Key Vault Secrets User and assign this role to the Server and optionally the BI containers.
Navigate to the key vaults Objects / Secrets and click on one of the secrets. Click on the secret's verison and notice the created Secrets in our vault have a Secret Identifier which is basically a URL that tells authorized Secrts Users where to find the Secret's value. Make a note of each Secret's Identifier.
We will now give the containers access to each of the secrets individually. Go back to Container Apps and navigate to your Server's/BI's Settings / Secrets. Click on Add and give the key a name, choose Key Vault reference for the type and fill in the secret URL with the Secret Identifier. For Managed Identity choose System assigned. After you've done this for each of the secrets let's use them for the environment variables.
The server container needs access to the DB, Cache, License key and SAML Certificate secrets.
The BI container needs access only to the DB secret.
Under Application / Containers in our server Container App hit Edit and deploy.
Here we have to do a couple of things for the:
Under Scale, choose the Min / Max replicas to best fit your use-case. Mine are:
Server: 2-10
Client: 1-2
BI: 1-4
Click on the container image's name under Container / Container Image.
Following applies only for the Server and BI containers - Here in Health probes we will provide two of them, a Liveness probe and a Startup probe
Liveness probe:
Path: /health-check
Initial delay seconds: 10
Period seconds: 10
Additional settings / Failure Threshhold: 2
Keep the rest default
Startup probe:
Path: /health-check
Initial delay seconds: 5
Period seconds: 3
Additional settings / Failure Threshhold: 20
Keep the rest default
Add a Scale rule for the Server Container and provide it these exact values (the Rule name can be anything). You can of course create your own values and types of scaling but I recommend using these ones as they've been proven to be most effective for mid-range loads.
Server Container Environment variables example:
Provide the Server container with the following environment variables
REDIS_URL
MONGO_DB_URI
LICENSE_KEY
CLIENT_URL (trailing '/#' is mandatory)
DB_TYPE
API_URL (the address of the server itself)
Optionally the following:
ALLOWLIST - to control what email domains can be used in Email/Password Auth
SAML_ENTRY_POINT - for SSO
SAML_ISSUER - for SSO
SAML_CERT - for SSO
SAML_CALLBACK_URL - for SSO
Client Container Environment variables example:
BI Container Environment variables example:
Your environment should be running now. It goes without saying that this is just one of numerous ways to structure your DecisionRules architecture on Azure.
Incorrectly following the steps listed above can result in your application not running properly. There are also a few things you might want to check after everything has been set up. Below is a list of known problems and checks you might want to know about.
Troubleshooting an error this broad can be difficult so we try to populate our software with clear error messages that should let you know what's wrong. To see them, go to your server container and navigate to Monitoring / Log stream ( You could also look up your container's log history in Monitoring / Logs ).
This is usually cause by the Client container's scaling settings. If they've been set improperly, no client containers might be live when you first try to access the site after a period of inactivity. Set the Client container's scaling to have a minimum of 1 replica running.
Azure Cache for Redis has a 10 minute idle timeout. This means that if the Cache hasn't been used in ten minutes it goes into sleep mode disconnecting itself. This can be prevented by setting the REDIS_PING_INTERVAL Environment variable. This variable is set in milliseconds and should be set to a time less than ten minutes. ( 300000 recommended )
If the Login via SSO button takes you to your provider but after inputting credentials you're getting an error there are two items you should check. Make sure the CLIENT_URL environment variable has a trailing '/#' in your server container. If the error persists check the SAML_CERT environment variable and it's correctness- It being one-line without ---BEGIN/END Certificate---.
To check whether the connection from your server container to your cache and database is private, navigate to Monitoring / Console in your server container and try to ping the name of your cache and database. If the connection is private this address name will resolve into a private IP address instead of a public one ( i.e. [10.x.x.x] or [172.x.x.x] or [192.x.x.x] )
Note: Adaptations might be required based on Azure updates. Always refer to the latest Microsoft documentation for current practices and configurations.
Please note this tutorial follows a naming convention similar to the one . This is not necessary but is considered best practice.
Read more in .
This part is the same as when deploying to Azure's Kubernetes Services. All of the necessary information on the provisioning itself and the cache's settings can be found in our .
This part is the same as when deploying to Azure's Kubernetes Services. All of the necessary information on the provisioning itself and the database's settings can be found in our .
This part is the same as when deploying to Azure's Kubernetes Services. All of the necessary information on the provisioning itself and the Enterprise applications' settings can be found in our .
All of the necessary information and more about Environment variables can be found on our .
