We are really excited to welcome our new wasmCloud maintainers - join us on Slack for the wider discussion, and to connect with our new maintainers - Github and wasmCloud handles linked below.
@ossfellow has been in our community calls and contributing many large features to wasmCloud like the platform helm chart, NATS KV store provider, and the NATS Object Store provider. He has also been serving as a great community lead and Slack troubleshooter.
@Adi sal is another long-term community member who has contributed many features/fixes and is constantly grabbing issues to work on.
@luk3ark built an entire platform on wasmCloud and regularly is pushing wasmCloud to its limits, catching bugs and race conditions while contributing fixes back to benefit the whole ecosystem.
markkovari is usually one of the first people in our Slack to respond to new community members and is invaluable as a leader in the community. He drops in often to take up issues that are on our roadmap, and has contributed quite a few PRs across the project.
@Jiaxiao Zhou (Mossaka) is a long-time Wasm community member who has contributed both to wasmCloud and is a WASI Cloud champion. He also has collaborated with other wasmCloud community members on the work for making Go awesome for Wasm projects.
Thanks to the community for your nominations and watch the recording to hear more from them (and an intro from Zhou!).
DEMO + DISCUSSION: Embedding wadm + wasmCloud in wash
This made it on the roadmap with a promise to spike on it. Check out the extensive discussion in the Github issue. Roman's feature suggests: if we have a Rust binary in wash the wamsCloud host is just a lightweight binary around the wasmCloud host library and so why don't we embed the wasmcloud host crate into wash?
As always, it's difficult to express the intricacies of this demo in notes so check out the recording below.
There's been a lot of discussion around the pros and cons of doing this (see the GitHub issue). A few considerations, for instance, Hyperkube includes all the Kubernetes CLIs - too big and perhaps not what we want wash to do,
We wanted to see what it might look like to embed wadm and wasmCloud into wash, The main difference is when we are setting up a process like wash dev in an ephemaeral environment, we stand up the local wasmCloud platform load up your component and its capabilities, you can do your hot reloading then close down when done.
What we do currently is download NATS, wadm and wasmCloud from GitHub releases (latest versions) and execute those binaries. We can clearly see what happens when we start and stop these binaries.
Follow along in the recording to see the process in action and hear the wider discussion.
We have noticed there is a growing appetite for benchmarking and so Taylor has created a new Helm chart in the repo for this purpose. We will also be using this to release standard benchmarking for wasmCloud relatively soon.
This is targeted at people who are running inside of Kubernetes, but technically it could be used for benchmarking wasmCloud deployments outside K8s and you're just using K8s for load-testing. This new chart is now available for the community to use - check out the README for detailed instructions and tips.
Simple process: in a normal cluster we can install a benchmaek from OCI - the one thing required is a test URL. Simple stack that is isolated from the rest of your stack.
Once we're up we're able to see where to see the logs and graphs. Test environment using K6 load-testing (Grafana) disrubutes a bunch of load to various locations. OTEL is capturing the data. We are able to rotate through our deployment in real time: run a benchmark and then review memory, CPU, pod usage, latency, failure and success rates over time.
There are multiple ways to configure depending on your use case. Taylor covers a lot of ground so check out the recording for the full demo and Q&A.
We had our Q1 planning session last week and the result of that can be seen in the Github project - we have a ton of great new features in the pipeline (including Taylor's Helm benchmarking work). Our docs pages have now also been updated. There is a nice retrospective and overview of the focus there.
Our high-level goals this quarter are stabiization, rigour, capability providers, SDKs, continued adoption of Wasm standards, and WASI P3 integration.
Our core and ongoing goals remain to enable developers to build applications with WebAssembly components without vendor lock-in and platform dependencies.
Our top level features are set and we have moved all our planning to Github projects. Aim: 1 item a week will get us close to achieving eveything on the roadmap.
wasmCloud Innovation Day recordings are being uploaded to the wasmCloud YouTube channel in the Platform Engineering playlist. We particularly recommend Bailey's AI and Machine Learning talk in which she describes the potential benefits of running AI/ML workloads with Ollama on wasmCloud.