This week marks a milestone for Intel and Snap, our open telemetry platform that allows you to easily collect, process, and publish telemetry data at scale. We are thrilled to announce the release of our 1.0 version of Snap, almost one year from when we launched our original beta.
When I look back on the past year, I am so proud of all that the team has accomplished. We took Snap from idea, to open source project, to 1.0 release, all while creating a vibrant community around it. The Snap team has spent the majority of 2016 working hard to improve the features in Snap – the library of plugins available, better extensibility for developing new plugins, and the testing infrastructure needed to maintain support for this production-ready release. We also created some valuable partnerships with companies like Netflix, Staples and Grafana, all of whom will be crucial in helping Snap continue to evolve.
A few of the highlights for Snap 1.0 release:
- Snap has gone from eight plugins at launch to over 70, with contributions from multiple partners
- We added support for intelligent metric selection through Dynamic Metrics
- We now officially support Snap plugins in multiple languages, and we released Snap plugin authoring libraries for C++, Python, and Golang
- Official testing and release process including package/automation support for RPM, Deb, Docker, Puppet, Chef, and more
- Our plugin API model moved to gRPC for better stability, speed, and streaming support
- A laundry list of new feature including distributed workflows, unified logging, CRON schedule support, build automation, and more
Real-time telemetry analysis provides insight into cloud and data center operations While many telemetry frameworks exist, without a consistent and flexible API it is tough (and expensive) to take advantage of telemetry's full value. That's where Snap comes in. Through a single API, it exposes measurements that provide a more comprehensive view of infrastructure capabilities, utilization, and events in real time — making full automation and orchestration of workloads across server, storage, and network resources easier and more intelligent.
The Snap team could not have accomplished a 1.0 launch without the feedback and support of Snap users and partners. The SRE team at Staples, for example, developed an open source monitoring platform with Snap that allowed them to consolidate their monitoring stack across Staples’ vast, heterogeneous environment. As an active contributor of the community, they’ve made significant contributions to the platform and written plugins of their own. They’re a great example of how an enterprise organization can benefit by putting telemetry data to work.
The 1.0 launch of Snap provides a stable foundation to continue our push to leverage hardware and software telemetry to improve cloud computing. On top of this foundation, our teams already have ongoing work on sophisticated Intel hardware health modeling, workload performance intelligence, and cluster scheduling insights.
While we are proud of the journey so far, we are more excited about what is to come and how we can leverage Intel's data center technologies with Snap to solve the complex cloud challenges of today. We are already working hard to expose Intel's 3D XPoint™, Resource Director Technology, networking accelerators, and FPGAs through telemetry and leverage this insight to drive a higher level of automation throughout the data center. Stay tuned in 2017 for some major Snap-enabled releases in these areas.
To learn more about Snap and get involved in the community, join our public Slack team.