The company’s Ark project is set as a utility for managing disaster recovery. Heptio said the focus was on Kubernetes cluster resources and persistent volumes, with a configurable way to backup and restore from a series of checkpoints.
Heptio said current methods of recovery from direct replicas of the underlying cluster state have not proven consistently successful. Ark can provide that consistency, coordinate volume “snapshotting” that keeps pod association, and allows for a single command to restore an entire cluster or subset.
Ark also allows increased portability of clusters and supports cross-cloud provider migration of Kubernetes objects. Initial cloud support includes Amazon Web Services (AWS), Microsoft Azure, and Google Cloud Platform.
Ark is being released in an “alpha stage” to garner feedback, with plans to add other cloud environments.
For those challenged by managing these now more portable clusters, Heptio’s Sonobuoy diagnostic tool can run a subset of open source configuration tests in Kubernetes or customer plugins. The tool is cluster agnostic, creating a readout of a cluster’s core characteristics regardless of setup.
Heptio said it created the tool to be customizable, noting it expects “that the types of tests users run to verify cluster interoperability will change over time.”
Sonobuoy will initially be curated by Heptio, but the company plans to “work with the community to find the right home for it.”
Heptio’s focus on Kubernetes is understandable. The company launched last year under the guise of making the platform more accessible to developers running apps on-premises or in the public cloud. Heptio Co-founders Craig McLuckie and Joe Beda were part of the initial Kubernetes team at Google.
Kubernetes has emerged as the most popular container orchestrator being used by enterprises. According to a survey conducted as part of a recent SDxCentral report on container and cloud orchestration, 64 percent of respondents said they were using Kubernetes, 36 percent said they were using Docker Swarm, and 18 percent said they were using Apache Mesos.
That same report also found that more than one-quarter of respondents that had not yet deployed containers cited a “lack of management and deployment tools,” and 24 percent said they “don’t know how to scale containers yet.”