Hey folks!
Today I'm delighted to announce the release of Suborbital Compute Beta-3. Compute is our WebAssembly extensibility and integrations platform which allows end-user code (written in various languages) to be run directly within any software system, giving users more control over the applications they rely on. One of the most important aspects of Compute is the developer experience it provides to developers writing these extensions and integrations within our partners’ apps. More than ever, developers expect fast build times and real-time feedback during the development cycle, and this release includes a number of enhancements to further improve developer experience.
Make it fast
First up in this release are faster function builds. JavaScript, TypeScript, Rust, and AssemblyScript functions now build up to 2x faster thanks to improvements made in our builder service. By taking advantage of language-specific tooling such as pnpm and Cargo vendoring, we can now build the functions written by users in a fraction of the time. This allows for faster iteration and testing, and it ties directly in with the next new feature...
Make it flow
Streaming build output. Build logs are now relayed to the function editor's console in real-time as the build is ongoing. This is a big improvement over Beta-2 where the entirety of a build’s logs would be displayed at the end of the build, which could leave you wondering what was happening behind the scenes. By adding a websocket connection between the function editor and the builder service, we’ve been able to give you a live view of the progress of our new zippy builds!
Make it clear
Next up we have better test reporting. When writing and testing a function in the code editor, you have always been able to send some test data into your function to test that it behaves correctly, but until now errors were not being reported into the console in a very helpful way. We would previously show an error code, but that’s about it. This release includes much more verbose output that shows error messages and function output to make it easier to understand how your functions are behaving when things go wrong.
Make it scale
Last on the list for this release is cloud bucket storage! This was a highly requested feature, and we’re happy to say that Compute can now use object storage such as S3 or Google cloud buckets for permanent storage! Until this point, Compute relied on Kubernetes persistent volumes to store user functions and other data, but now we gain the fast and unlimited storage capacity of cloud buckets to make Compute more flexible. This also means that the builder service is more scalable, as it does not need to be co-located on a single node within your cluster.
We're continuously working to improve the developer experience of Suborbital Compute, and we appreciate your feedback! Reach out to us at team@suborbital.dev or join our Discord server to get in touch with us, and please don’t hesitate to ask for a hands-on demo. The coming months will include some very exciting announcements surrounding Compute, so follow us on Twitter and stay tuned.