Transcript: wasmCloud Q1 Roadmap, WASI Preview 3 & the Wasm Component Model
wasmCloud Weekly Community Call — Wed, Mar 26, 2025 · 30 minutes
Speakers: Brooks Townsend, Victor Adossi, ossfellow
Transcript
Brooks Townsend 05:27
Welcome to wasmCloud Wednesday for Wednesday, March 26. This is the last community meeting of the quarter. The reason that's significant is because we do quarterly roadmap planning for wasmCloud. So today I want to take a look at the roadmap. We have a couple of agenda items, but by and large I foresee this meeting being pretty quick. We should have plenty of time at the end to talk through additional agenda items if necessary.
As we look toward the next week — well, tomorrow is going to be WASM I/O, and then next week is going to be QCon and KubeCon + CloudNativeCon EU in London. All of us are preparing talks, you know, last touches on demos, so we're pretty heads down. No demos or super-big discussion items today, but just to share the community meeting agenda: I wanted to leave a little time to take a look at the wasi-config update discussion item from last week, which we just rolled over; look ahead to WASM I/O and KubeCon Europe; and then check in on the roadmap. I think we'll do this in the opposite order as proposed — because these are unordered lists, right? We can do them in whatever order we want.
So taking a look at the roadmap — we've done this kind of check-in a couple of times so far in the quarter, and I think we've progressed fairly well. We knew during the planning phase — I mean, this was the first thing we said right when we got off that call — "wow, we planned a lot of stuff on this roadmap, we're going to see how it goes," because we are operating at the bleeding edge of WebAssembly. There are many, many things that we end up needing to adapt to, work with, and help implement, and it's kind of hard to flesh out a full roadmap of all the stuff we know we're going to do in a specific quarter. So that spiel is just a little bit of prelude.
But if we take it down to the bottom, we have completed a good couple of tasks on this roadmap. Open-sourcing and using the benchmarking Helm charts, so that you can profile and benchmark your wasmCloud applications in a reliable way. Combining the wash CLI and wash lib removed a point of friction in releases that was essentially tripping us up and making it harder to release, which is now done thanks to the great help of Ahmed. Our maintainers Taylor and Roman have come in with a ton of improvements to releasing secrets — that's KV — and organizing our crates and CI to only build providers and wasmCloud when we need to.
We also have a lot of improvements that came in for our documentation, especially around using some special GitHub features — I say "special," it's just a GitHub feature — to use conventional commits to generate release notes, so that when we cut a release of wasmCloud or wash they're nice. They're nicely separated by markdown headings that we've determined, and then any breaking changes and things like that are surfaced all the way up to the top.
To take a look at some of the things we have in progress — these are action items, or work items, that we've started on. They may or may not have a PR out there, but are essentially in progress. This one, emitting the error count when the logging buffer is full: this is around when we do our tracing emitting. If we exceed the internal Rust buffer for events, then essentially we lose that trace event, which is by design — we don't want to block a component on tracing. But this one actually has a PR open, thank you Rob Marvin for opening that. So we just need to push it through the review queue and get it wrapped up.
Same thing with a cron-job provider. De opened a PR with a cron-job provider. We've actually had a couple of other folks in the community also design cron-job providers — I know I've seen a couple. So it would be good to collaborate on the right design. Aditya is a maintainer of wasmCloud and has reached out that this is essentially a very busy time for him personally — and open source a lot of the time, you know, not everybody is able to spend their work time on it. So this one is in an open PR pending some discussion, but really is moving along well, so I'm not worried about it at all.
Bundling wasmCloud and wash has stalled a little bit. This was my task — I really want to do it, it just hasn't come up to the top of the priority list. So I expect to probably hack on this maybe next week during conference time when I'm procrastinating something else. This one has been proven possible, it just needs a little prioritization to get it done.
WASI P3 tracking is probably the biggest issue, or biggest card, on this roadmap. And there's a significant couple of maintainers who would normally be working off of this roadmap that are actually pushing across and helping contribute to WASI P3. I want to especially call out Roman, who has been implementing a lot of WASI P3 in Wasmtime — the actual async versions of interfaces, the new async versions of interfaces, and the P3 variations, like with WASI HTTP, actually implementing them in Wasmtime. It's being done by Roman and others in the community. I'm not actually sure who exactly is contributing the features, but I'm sure that Alex Crichton and Joel Dice from Fermyon are putting in a good bit of contributions too. So I'm not intending to say that we're doing the whole thing — I just want to call out the wasmCloud maintainers who are doing awesome work in the standards space.
And then I also want to call out Victor, who has been implementing a ton on the JCO side of things. I believe he's working all the way across the JavaScript ecosystem, and on P3 in general — there's been a ton of work there. Victor, I think you're on the call. I'm not sure if you wanted to say anything more there, but I just wanted to call out and give a huge thanks to the folks working on the upstream WebAssembly standards too. I think it's very important for us to be contributing as much as we can to the standard. So that's likely a big reason that some of these items are going to slip over to the next roadmap, or maybe just weren't as important — because getting P3 out is a very important task.
Clap autocomplete is also another one that's in a PR that I think we just need to drive across the line. Dude is working on the built-in HTTP client capability provider, and said he'd be putting out a draft PR for this pretty soon, which is great. And then Masood posted a good bit of information around this RFC for SBOMs and build provenance. I believe the conclusion there essentially was: hey, this would probably further complicate CI and wouldn't give us much tangible benefit right now, so it's probably not worth doing. I think we still need to talk about it, but I really appreciate the investigation into it, Masood — we should probably just bring it up again on another community meeting to talk all the way through it. Victor left a comment — Alex, Luke Wagner, I think you're talking about Luke Wagner, and Roman and others, everybody contributing on the standards side.
Okay, so those are the things that are in progress and completed. The majority of the roadmap we have at least started on, which is great. The remaining improvements that are "ready for work" are actually likely really quick tasks. If we got together and did a hackathon, this would be a very quick couple of things to knock out. So in the name of getting things completed on the roadmap, I think some of us maintainers should be able to just get through and tackle these — but they're still up for grabs for anybody looking to contribute.
Other than that, the couple of things that have been left in triage intentionally are not quite ready for taking by anyone. Usually it requires some more maintainer or offline discussion to figure out the way we want to solve these problems. I think this is a good lesson to learn for our next roadmap planning: the items we bring under the roadmap should be well scoped and ready to take at the time of roadmap planning wherever possible. And that's totally fine — like I mentioned in past roadmapping sessions, I'm not a Jira scrum master trying to whip everybody into shape across the community. I really just want to show the broader trend, the bigger issues, some of the things we really want to get done, in a place that everybody can see them and that we can contribute to.
I don't think I have anything else to say on the roadmap, other than: if you're taking a look at wasmCloud, and you're really interested in contributing to this super-cool project, take a look at this "improvement, ready for work" pane. I think it's a great place to get started for contributing. Victor, go ahead.
Victor Adossi 17:19
Hey, hey Brooks. I just wanted to note for everybody who's watching and part of the community that is also filing issues: we've had some really nice usability issues and things people ran into come in, and that's been really helpful to just go through and make sure we have really good docs, and then hit bugs and make sure things are easier to use. So you don't have to wait to find a specific issue to work on — if you just start going through and you're building something, those are honestly some of our favorite issues to get, because then we can find edges that are a little sharp.
Brooks Townsend 18:03
Yeah, awesome, awesome call-out, Victor. Thank you.
Okay, so the only other thing I wanted to share about the roadmap — which, by the way, this is totally open for anybody — is to take a look at the wasmCloud Dev stats instance, which is provided for us by the CNCF. It just kind of lets us take a look at all kinds of statistics for a certain time period for the wasmCloud org. You can see, over the last quarter, wasmCloud may have actually merged 335 PRs — an average of a little under four a day. So I think this speaks really well to the idea that we may not take everything off of the roadmap and get it done, but we are certainly busting our butts. We're getting tons of contributions from people in the community, which is really awesome to see.
I always forget which one is the best to look at here. This was from three years ago, but you can see the number of contributors and contributions just steadily upticking over time, which is really awesome. I think this is the window of the new year — so holding really, really steady on the contributions and contributors front, which is awesome. I'll try to publish a couple more of these; it's fun to look through some of these graphs as well.
Okay, well, we have certainly been putting in the work, and I'm looking forward to our next road-mapping session. The only thing is that KubeCon + CloudNativeCon is next week, so during wasmCloud Wednesday, myself and many others are likely to be at the conference itself, at the booth. So next week's community call will probably be pretty short — from the conference floor, one of us yelling over the noise of the crowd to tell you how it's going. So it probably won't be the best time to do our road-mapping planning next week. I'd like to push that at least to the week after KubeCon. We don't typically worry about it too much if we miss a week or so.
As far as the other things — Cosmonic has a super-healthy, helpful blog on their site for "Cosmonic at WASM I/O and QCon," and this just happens to include many of the wasmCloud maintainer talks and things we're going to be involved in. It highlights some of the wasmCloud things as well. So for our next agenda item, we wanted to look ahead at WASM I/O and KubeCon.
So here's what's going on. In Barcelona, which will be tomorrow and the day after — I know that our very own Bailey Hayes is already over there, and I think Liam is on the way, going to be flying there soon, or is over there. Actually, can't remember — Liam is moving around a lot these days. But they are going to be doing a talk at WASM I/O. And then I believe — I don't know if I have anybody here from Couchbase — but I believe Couchbase is doing a workshop at WASM I/O which includes wasmCloud. So both will be awesome sessions to catch. I personally won't be there, sad face, but if you're looking to connect with anybody on the wasmCloud side when you're in Barcelona, you should hit up Bailey and Liam.
And then at QCon, we have quite a few things happening. Taylor, on the wasmCloud side, is going to be doing a lightning talk with TAG Runtime — some of our involvement in the CNCF Wasm ecosystem. Jonas, who contributed the workload identity feature to wasmCloud, is going to be talking about Linkerd and WebAssembly and containers. Then Taylor again, and David Justice from Microsoft, are going to be talking — one of my favorite talk titles so far this year — "Was Am I Right, or Was Am I Wrong? An Overview of the Wasm Ecosystem." Super fun. I have a quick talk — I guess there's a normal talk — I have a talk, "Wasm Whiplash: wasmCloud's Wild Ride to Standards," which I hope will be fun. It's a five-year retrospective on wasmCloud as a project, how we got to WASI P2, all kinds of fun stuff. Hopefully it'll be as enjoyable for everybody attending as it is therapeutic for me to talk about everything on my mind about the project — all good things, all good things.
Then one that I'm sure many folks are looking forward to: "SPIFFE in Practice for WebAssembly Workloads," which is going to be wasmCloud maintainer Colin Murphy talking about workload identity. And then finally, Liam actually has a talk in the platform engineering track, in the main conference — we're going to see and talk about "Can You Maintain 1,000 Apps? wasmCloud and Kubernetes." So there should be a lot of production learnings in that talk — the things that we've seen wasmCloud do after quite a couple of years of running it. So it should be really fun. I've linked this blog in the Zoom chat — it's on the Cosmonic blog. Got quite a couple of talks coming up. So I hope to see you all there. Anybody who's going to be in London, please let us know, we can go grab a coffee, grab a pint, whatever.
And yeah, I think that's it. Our only other agenda item that was on the list for today was an update — a PR, or I think it was an issue, that was filed on the wasi-config side. But I do see that Taylor isn't here, couldn't make it, had a conflict. And I'd like to get his perspective — I'm happy to read out the issue, but I'd like to have his perspective, so we'll bump that one one more week. As I mentioned, a lot of the folks on the wasmCloud side are pretty heads-down shipping features and talk slides for next week. So that's actually all I had for the community meeting, but we do still have time if folks have other things they wanted to bring up — from wasmCloud, WebAssembly, the broader Wasm ecosystem. So I'll shut up for a second, and feel free to raise your hand if you do. Soon, go for it.
ossfellow 25:54
Hi Brooks, can you hear me? Today I'm on my mobile — first I was on YouTube and didn't have a different experience. Quickly on that issue: do you want to leave it till we talk about it in detail, or do you want to discuss it now? Because either way I'm okay. I know what I have done — I have spent a lot of time on that ticket, and really it's weighing the options, you know, which way we want to go, and what the pros and cons of each are.
Brooks Townsend 26:39
Yeah, well, we could talk about it today. I want to make sure that some of the other folks you've been talking to, like Jonas and everything, can weigh in.
ossfellow 26:56
And that's why I'm saying: if you want to wait till basically KubeCon and those are over, and then when we're talking about the roadmap, potentially we can talk about what's left and what we're going to do with it.
Brooks Townsend 27:13
Yeah, that sounds good. I am good with that — good with leaving it there, just to double-check: by leaving it there, there's no more work you're doing on that right now, right? We're just waiting to talk about it?
ossfellow 27:31
Right. And making that decision — that if we want to go ahead and implement it, or say that it's not worth it, and we want to wait until we have, let's say —
Brooks Townsend 27:49
Yeah, okay. That sounds great. I think it's really worth the discussion. I don't want us to feel like we have to go through with features if we investigate them, and then —
ossfellow 28:02
No. And the fact is that I have run the jobs — there are logs, there are outputs, there are reports, there are actually attestations. Everything is there for the demo purpose. So when you're doing the review and making or deciding or debating what to do, you have material to look into and say, "okay, we get this and we don't get that, we can be like this and we don't like that."
Brooks Townsend 28:34
Yeah, totally makes sense. I think it makes a lot of sense to bring this up after KubeCon, when we're thinking about what we want to do for the next quarter.
ossfellow 28:45
Sounds good. Thank you.
Brooks Townsend 28:55
All right. Well, I think that may be it for today — a little quick community meeting. As a reminder — and I did say this a little earlier — I think next week will be a quick community meeting from the conference floor. So I'll be yelling at you, and it should be a good time. If you're in London, come be on the community call in person, it could be kind of fun. Other than that, look forward to seeing a bunch of you next week, and then picking up the community call, and I guess see you back here in two weeks too for our regular community call to talk roadmap. All right, thanks everybody. Have a great rest of your day, safe travels to everybody going to the conference, and have an awesome Wasm day.