The KEDA project has hit v1.0, six months after Microsoft and Red Hat first unveiled their bid to deliver event-driven scale capabilities for container workloads
The one-time arch rivals announced the project back in May. At the time, principal PM manager, Azure Serverless, Jeff Hollan said that scaling in Kubernetes is “reactive” and the aim of KEDA was “to bring the benefits of event-driven architectures and the productivity of functions to Kubernetes”.
According to Red Hat’s announcement this week, KEDA allows Kubernetes to “scale applications using a pull-based mechanism, and enable Kubernetes to scale apps to and from zero.”
Hollan wrote separately: “Kubernetes provides a powerful container orchestration platform, but by default will only scale based on system metrics like CPU and memory. This means there can be a significant delay in a system’s ability to respond to events.”
“With KEDA installed,” he continued, “Kubernetes now can know how many messages are arriving, and use that information to start to scale an app even before the CPU starts to rise.”
The 1.0 release supports a number of different scalers, “including Apache Kafka, Azure Event Hub, Azure Service Bus, RabbitMQ, Prometheus, AWS SQS and others”.
Hollan added, “KEDA also has an extensible gRPC contract so other scalers can be dynamically discovered to add additional capabilities.” KEDA can also run alongside Virtual Kubelet and Azure Kubernetes Service Virtual Nodes
Red Hat said, “Our goal is to continue making KEDA a first class citizen in OpenShift, easy to consume and install, so you won’t need to be an expert and can focus on your application, freeing developers from infrastructure concerns.”
“Through KEDA, Red Hat can enable Azure Functions as a workload for the platform, available as a developer preview on OpenShift, delivering a hybrid FaaS solution that can consume events from Azure Services as well as other additional event sources.”
At first glance, it might sound like KEDA is playing in the same space as Knative, the Google-backed component spec, which amongst other things promises to “Run serverless containers on Kubernetes with ease” by handling eventing, and taking care of networking, revision tracking and “autoscaling (even to zero)”.
Red Hat said, “We see Knative and KEDA as complementary projects, where one can address the needs for push-based requests while the other is focused on pull-based auto-scaling.”
In other Azure/Kubernetes related news, Microsoft said it was supporting confidential computing for Kubernetes workloads, building on Intel Software Guard Extensions (SGX). SGX works by creating secure “hardware enclaves” – chunks of encrypted memory – and Microsoft said the technology enabled a number of previously unrealisable use-cases.
“Customers in regulated industries can now collaborate together using sensitive partner or customers data to detect fraud scenarios without giving the other party visibility into that data,” it said and “customers can perform mission critical payment processing in secure enclaves.”