Scaling dynamically with scale rules
With Azure Container Apps, scale rules can be defined based on concurrent HTTP requests, concurrent TCP requests, or custom rules. With custom rules, scaling can be based on CPU, memory, or many events based on different data sources.
A microservice isn’t necessarily triggered based on HTTP requests. The service can also be triggered asynchronously, such as when a message arrives in a queue (for example, using Azure Storage Queue or Azure Service Bus) or when events occur (for example, using Azure Event Hub or Apache Kafka).
Azure Container Apps scale rules are based on Kubernetes Event-driven Autoscaling (KEDA), which offers a large list of scalers. You can find the full list at https://keda.sh.
When using a KEDA scaler with the Azure Service Bus queue, you can specify how many messages should be in the queue when another replica should be started. What’s common with all the KEDA scalers is the configuration of the polling...
