Scaling Dapr on Kubernetes
In the world of monolithic architectures, the compute and memory resources that are available to an application are constrained by the hosts that it operates on; that is, VMs or physical nodes. For such applications, it becomes an extraordinary challenge to distribute requests and jobs evenly between multiple hosts. They often resort to an active/passive mode in which only a portion of the allocated resources benefit the application, while the rest are passively sitting idle, waiting for the active environment to fail so that they can switch from their passive role to an active one.
The following diagram depicts the challenges of scaling monolithic applications:
Figure 10.1 – Scaling monolithic applications
Given these factors, to respond to an increase in client requests, which translates to a demand for more computing resources, the response is often to scale up the resources. This can be done by substituting the hosts...