Knative is all grown up
Ready for widespread production use?
Knative, an open-source project that makes it easier to run containerised applications on Kubernetes, has officially “graduated” as a CNCF project.
That signals Knative’s “readiness for widespread production use,” said the Linux Foundation-backed CNCF – whilst stripping out the heavy lifting of autoscaling, routing, and event delivery workloads on Kubernetes.
With serverless at its heart, Knative lets users autoscale services to high volumes if needed or scale them down to zero – unlike Kubernetes, where you need at least one instance of a “pod” of compute running at all times.
Platform Engineering: Lessons, queries from the First Hype Cycle
“Knative’s graduation reflects the maturity of serverless technology in the Kubernetes and CNCF ecosystem,” said Chris Aniszczyk, CTO, CNCF.
He added this week: “The project has built a strong contributor base, gained trust from end users and continues to evolve with integrations that address needs, from newer AI workloads to interoperability.”
See also: Solo’s CEO on the new “ambient” Istio
“Knative fills several gaps in the cloud native ecosystem as an easy on-ramp to Kubernetes, with Knative’s eventing acting as the missing skeleton for connecting events to reactions,” said Knative’s co-founder, Evan Anderson, in a canned statement, adding: “It's gratifying to see the vision that inspired Knative realized and adopted across the CNCF.”
The project relies on and contributes to the CloudEvents specification for interoperable events, integrates with Buildpacks in the Knative Functions component, and shares base packages with Tekton, a project that originated from Knative’s early build system, the Linux Foundation said.
The Knative project roadmap includes Knative Eventing to bridge synchronous and asynchronous workloads and enable the likes of MCP clients or legacy applications, to communicate with applications built around an Event Broker. Additionally, Eventing now integrates with Apache Camel Kamelets—bringing new event sources into the ecosystem.
The project has also switched to OpenTelemetry for metrics and tracing, for users to send traces to their preferred observability provider.
Knative contributors also suggest that it is an ideal platform on which to build event-driven architectures on Kubernetes for AI applications. Ali Ok, Principal Software Engineer at Red Hat, recently shared some architectural details on that approach here for those interested.