Open Feature Project, 13 Oct 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: OpenFeature - Project meeting, October 13th, 2022

Description

meeting notes: https://docs.google.com/document/d/1pp6t2giTcdEdVAri_2B1Z6Mv8mHhvtZT1AmkPV9K7xQ/
OpenFeature website: https://open-feature.github.io/

A

B

B

B

In all the meetings in the last two years, as ever, asked me why I have a street uh road sign in in my room.

C

uh I I had I was more the power rack that interested me to be honest,.

B

Oh, but that's that's, I, have the most famous gym of lockdown Within dynatrace. Well,.

C

No, what okay! This is a story. I, don't know about. I need to hear it famous.

B

Oh because I did, uh over the last almost two years, the open source talks within dynatrace and like having 200 people connected. They would see the see the power rack in the back like like what what room is that was like it's, where I do my hard work.

C

Yeah yeah there you go.

D

Hey everyone morning, Pete good afternoon, Ben Thomas see you guys um just a reminder. If you have any topics, feel free to dump them in the uh meeting notes here, I'll just wait a minute or two and go ahead and kick it off.

E

Hey I've got a question when, um if I think of a topic like ahead of time, do I just am I supposed to put it in the open items or is that? Is there a different place? To put it.

D

Just throw it on the agenda, just talk anywhere and usually just add. Your name at the beginning uh makes it easy, because I can make sure I can call on you yeah.

B

Yeah I, usually not always, it depends on sometimes on my on my um forgetfulness but usually I, create the the let's say the agenda slot for the next meeting at least a week in advance, okay, but yeah. If it isn't, there feel free to copy and paste yeah there's not much.

E

Magic to it, it's kind of funny, it's one of those things that is probably is presumably pretty straightforward, but I'm like I, don't know if I'm am I allowed to touch. This am I allowed to accomplish.

D

Maybe we should be a little yeah, add a little snippet or something above and just say. Like I, don't know, add your agenda items if you want so because yeah I feel like we get people to add it during the meetings, but not necessarily ahead of time. So maybe there is uh some hesitate, hesitation there so cool. um If you don't mind, just add your name uh to the participants list on there um and then I suppose we could probably just go ahead and get going. uh This guy is still good to do scribe.

D

Work today, perfect. Thank you. Any volunteers for scribe work next time. I would.

B

Have asked because it's during kubecon, oh.

D

um Okay, um should we I mean I, don't know if I'll be able to be there I think Todd will also not be able to be there very likely. Given the timing, I think the the booth actually opens at 10 30 on on a Thursday or on the Thursday that week, so we.

E

Can either be like a live Edition like from the company from the booth.

D

um I, don't know, maybe uh so I guess we have a couple options. We can uh try to make something work. um We can move it to the week after or we could just uh you know, kind of cancel it that week and just do something. uh You know a kind of a recap at our normal schedule. The week you know two weeks later, I.

B

Don't know, maybe we can still make it work. We just have one person that stays and covers the booth and.

D

Okay: let's try to make it work. Yeah I mean it would be nice to provide an update. It would be a Thursday into kubecon, so we'd already have a few days there and hopefully some some updates and feedback from from the attendees there. So that sounds good, uh so Pete we're gonna pull pencily in there. Unless you decide to come to keep con last minute tickets.

E

D

Only like 1800 bucks.

E

Yeah I'll ask my boss: oh no, he said no okay.

D

Fair enough all right, um perfect morning, Justin all right, uh so I guess the first topic is kubecon. uh Just a reminder: I hate to keep teasing it, especially for those that that aren't attending, but um I do have like a quick sign up Forum there. uh So make sure that you add yourself, um I'm also wondering if maybe we just want to create a slack channel for those that will be attending. So it's easier to you know, communicate, and then we just archive it when we're done.

D

um I think that really only applies to a couple people on the call. um So if you think that's a good idea, just let me know I'd be easy enough to do and we can go ahead and uh create it and then archive it uh and then the last thing real quick. If you missed my message in the slack Channel, if you are going to be a kubecon there's a cube Cruise, uh which is on the Detroit River um and it looks it looks like it should be. Nice I would definitely recommend dressing.

D

Warmly um Detroit in end of October could either be beautiful or snowing so yeah pack in plan. Accordingly.

D

uh Next thing is just the demo application update, so we kind of teased it at the last community meeting, but now that work is essentially finished, if you haven't already definitely check out the playground repo, um it basically is a fully. You know self-contained like Docker compose files now and you can just run Dr compose.

D

It comes with uh the environment, variable the flag, D and uh go feature flag out of the box, and then, if you, you know, register any vendor, you can just add that their access key as an environment variable and it will show up in the list- and you can just select it um and then try to make it. You know update the readme to show how to set up each of the different vendors and then have like a little. You know, marketing blurb, for for each vendor.

D

On kind of you know what their value prop is: there's also kind of a story around. You know why you know fiber decides that they want to use feature flags and what the value is there. So if you have any feedback on that, certainly let me know um there is a few like I. Would say remaining tasks around like talking about the open, Telemetry integration and a few other things, but the basics are all in place um and so yeah definitely take a look.

D

Let me know uh and and appreciate, Todd's help making it look a little bit nicer. uh It's certainly still like I I, don't think fiber is going to make or Fiber is going to make a ton of money here. But you know it does look a little bit a little bit nicer um yeah. So so that's uh that's definitely worth checking out another thing: that's kind of cool.

D

um There was a Community member uh that helped work on a killer, Coda demo, it's something that I was not familiar with, but it was a kind of a cool tool that basically spins up uh an environment for you in your browser and then starts like the repo automatically and just exposes all the ports and stuff like that.

D

So basically, as long as you have an account which you can link link to um your you know your GitHub account or something it spins up the demo and just exposes the ports, and you can start playing with it without having to to pay for infrastructure, no idea how they're going to stay in business, but I guess we'll just rip through their their venture capitalist money until they uh they run out and we'll pivot. So pretty cool definitely check it out.

D

If we want to add more uh more demos scenarios whatever this may be a cool spot to do it, it looks really simple to add stuff, and there is a repo in our org now for just adding new scenarios, so you know take a look uh but I think it has some opportunities and I was definitely impressed when I went through it for the first time, because it just kind of works, so you don't have to you know mess around with even having like Docker and stuff set up on your computer, so kind of cool uh any questions or feedback on any of those, um and then next one is just a really really quick.

D

Hotel update this is the uh pull request that will never get approved. It feels like it's been open since, like you know, beginning of the year, uh it has gained a little bit of traction recently, um just based on you know. We do have four approvals, which is technically sufficient, but there's a few people uh in the hotel Community, that's a bit concerned about using spans versus events versus attaching it to an existing span. I've experimented with different versions of that. So you know it's.

D

It's kind of on me now to to lay out the pros and cons of all the different approaches: I'm leaning towards an event at this point, um but I just need to kind of spell it out and and I'll probably provide the different options, maybe in the demo environment. So we can see them all side by side by side and hopefully get some some kind of approval. I. Think most people from the hotel Community are okay with something there's just not been it. You know people accepting in the existing approach unanimously, I suppose so.

D

That's kind of the current state I'm really hoping we can get something uh worst case. You know, I I have some conversations that keep calm with the the team and just see what we can do here, because uh I I think it's it.

D

Shouldn't have been that controversial, uh so I'm hoping we can get something in soon um next thing, uh I'm not even sure if most people even realize this was an issue, but there was an issue with our our org, where GitHub Pages weren't working, it was trying to use openfeature.dev and not not working properly. So any kind of like repo specific GitHub page was not uh not loading that has been resolved. It was a very bizarre issue.

D

It took me a lot of time, but uh we're good to go a quick summary in case you're ever managing a GitHub auric. If you have a uh what is it in our case, it was like an openfeature dot, GitHub IO uh repo. In our case it was archived even um and then that was set up with a GitHub page. It messed up all of the other repos. So I had to rip out that configuration from that archived repo and everything worked as intended. So all good there.

D

um The way you can see I guess uh the result of that is in our JS repo we're publishing like the uh like a automatic like typescript generated, uh you know page, so you can kind of view all of the the public methods and interfaces, and things like that. I can pull that up. I'll link it in the chat uh in a few minutes, because I think that's basically the last topic that I have on the list right now. So if there's no questions about any of those I can hand it off to Skye.

D

Nothing cool uh Skye go ahead. Please.

A

um So, a week or so ago, Ryan Lam raised an issue in the net SDK, with concerns of threat safety, specifically around the evaluation context, which he had found to be mutable, which was concerning because say, if someone were to create an evaluation context, object, pass it to a flag evaluation and then mutate it in some other thread somewhere, and then this could have undesirable consequences.

A

Obviously, um so this then led to investigating whether this was an issue in other places, and we found that it was in the go SDK and to Java SDK as well I believe so breaking changes were needed and the way this was resolved was to create a well in the go case, at least to create a Constructor um and that effectively means to the instructor is immutable now, and we we also added in some some blocks at the client and evaluation level, to make sure that any time that a provider is updated, um the ongoing evaluations just keep using the provider.

A

That was at the start of the transaction, um and this means that once you update a provider, it only takes effect from any transactions to begin. Beyond that point, to avoid these uh undesirable race conditions.

A

That's it cool. Thank.

D

You um I guess one last thing about go uh the the test harness was there's a pull request open for that as well, so that will be like the end-to-end tests, uh which I think is, is basically one thing that we wanted to get in before we had the confidence that that we're kind of good to go with this um and all the breaking changes for Java go and Dot net have already been published, so uh I think we're in pretty good spot there that those are the last like known breaking changes that we wanted to get in prior to kubecon um and so I think we're in a pretty good spot.

D

E

So do we need to do anything with the spec to kind of does that? Does this change anything in the semantics of the spec or is it just kind of an implementation detail for the for the for certain languages.

C

I think it's more of an implementation detail for certain languages. I mean JavaScript, JavaScript doesn't even really have threading concerns, um so yeah there is. There is kind of the The Wider, the slightly wider issue of um changing a provider over the course of an evaluation which um potentially like that that could be applicable in the case of JavaScript, but.

E

C

Like um you know, uh specifically like reordering due to lack of walking and that type of thing are only problems and threaded languages, so.

E

C

E

Of making of the valuation context in mutual isn't, is that saying there's either just under like undefined in this spec, whether it's mutable or not? Well, would it be something we should say is encouraged so that.

F

E

Someone's doing this in another language, they don't run into the same issue.

C

Yeah I'm hesitant because I mean to me the biggest issue with the mutability around things like um things like the context or even the collections like hooks, for example, the biggest Financial issue is if it was if, while it was being enumerated, it's modified that usually causes an exception in languages like Java and c-sharp, for instance, so like these Primitives, when you're iterating over them and then they're mutated, you'll actually get an exception. Throne and, and that can only happen in a situation where you have threading.

C

So again, it's like I feel like it's not applicable to all languages and it's a concern specific to those languages, um I'm not against like encouraging immutability, uh but it seems like it's only kind of it seems like it's only like really an emergency or or like something we'd want to handle for sure in sdks and languages that are impacted with those kinds of issues but by those kinds of issues. But that's my take I, don't know if anyone disagrees. I wouldn't I wouldn't be opposed to adding some kind of encouragement to make.

C

You know to consider immutability consider threading. We could even do it in a non-normative section in the spec of a felican.

E

Yeah I bet, yes, I, guess you're, saying it'd be more like kind of like a you know, advice to implementers, rather than like a requirement in this pack.

D

And maybe that is the way to do it, because that's something that wouldn't require massive effort, and it's probably worth at least mentioning that, like you want to think about threat safety, because otherwise you know if we would have done that from the beginning. You know we probably could have avoided at least a few of the issues um and it's it's kind of one of those lesson Learners to some extent,.

E

D

Some yeah, maybe just adding a quick little blurb on on the stock, could be helpful.

G

But what happens if you? um What happens if you want to change the uh some data in the context like as part of us, you know if someone updates their profile, and you know that that logically modifies, like.

E

During like, during a request, it makes someone an admin, for example me, but maybe a realistic scenario.

C

Foreign yeah, so there's there's no there's no problem with changing the context at at any level. You just can't do it by mutating the actual object, so you have to construct. Oh.

G

Okay, so you'd have to prove like instantiate a new context or something.

C

In the in the places where it's implemented- immutably, yes, yeah right.

E

Like at the point that you're handing it off to the to a flag evaluation, you're, basically like freezing it at that point, right.

G

Okay, so okay yeah, which makes sense.

A

Although the beforehoops can return a new evaluation context, so that is one way of then changing it. Mid evaluation.

D

But you can't modify it in the before hook. You have to return a new one, correct.

A

D

So and that's optional, so.

C

Just just I saw Justin and mute his mic a few times. It is something you.

H

Want to say sorry, no worries, I um I, think that, like how we handle the threat safety issues are implementation, dependent, I, think it makes sense that we would say in in a normative section say you should consider threat safety in your implementation and then in a non-normative section say something along the lines of like okay in other sdks. Here are the areas in which thread safety was a concern. Here are the scenarios you should care about, yeah.

D

That makes sense Todd. Is that something that you can uh work on? Please.

C

Yeah I can I can open a I, can open a PR um and I I like Justin's overview. We can. We can just kind of touch on it uh and vaguely, even in the normative section and then yeah we can expand on it. If you feel like it.

D

Perfect. Thank you uh anything else on on your section, Todd I! Guess you have you even really gotten into the hardening of the the spec.

C

Okay yeah, so um my my section is uh is pretty short and sweet, so we talked before um about our kind of our goals for uh our goals for hardening the sdks uh I'll find a link for that after I'm done speaking and add it there, but um I do have a pull request open uh to actually add the hardening marker um to a couple sections.

C

um This pull request also makes a few changes, just in terms of structuring the uh a couple documents a little bit better, there's no actual, there's, no actual functional changes, um they're all editorial. uh The thing is: I, do renumber some sections specifically in the provider, the provider was the only section we had that. Wasn't that didn't have like three decimal places for every requirement um and it made it a little bit weird and it also made it uh meant that it wasn't quite quite so forward compatible for adding new sections.

C

um So I I changed that so basically some section numbers some some requirement numbers changed that can be a bit irritating, especially for like things like the Java SDK, where there's um test Ketone sections I totally don't mind being the one to to fix those things, um but yeah the the I think the numbering is better I, also added section numbers to every single heading, because we want to be able to add these uh stability markers to particular sections and if sections aren't numbered I feel like it's kind of vague, um where exactly that uh stability marker applies to where it begins and ends.

C

So all of that's kind of captured in that PR feel free to take a look at that, um but yeah the the important part. The really important part is that the pr marks um the evaluation, API and the provider as hardening. So as per our earlier discussion and the other sections in the spec outlining this. That means that breaking changes from this point on to those sections require approval by the technical steering committee.

C

We don't anticipate, there's going to be any, but the whole idea is that it's going to give confidence to provider implementers and application authors that these areas aren't going to. You know change wildly and that we do. We are at the point where we can say: they're stable, for use in production systems.

C

Yeah, so so that's what that's, what the idea is of that. So take a look at the pr and let me know what you think I think: that's it for that section.

D

Perfect thanks Todd all right, Pete um yeah, uh some surfacing issue URL and other metadata.

E

Yeah I'm kind of just uh the messenger here, so the someone in the open feature slack was asking about this and I kind of figured. It wouldn't hurt to bring it up in the meeting. So there's a link in the in the show notes: there's a link in the dock here to um the discussion in stack, which I'm now following to refresh my my own memory.

E

So this is someone who I think works at last year and and was asking if there's a way to like kind of like have a uh a field in the flag that would essentially have like an issue URL or something like that, so that if, if you're a developer and you're looking at the flag, you have a way of I.

E

From my my understanding of what they're saying, do you have a way of kind of like going to like a page that describes the reason for the flag being there or like the states of the flag like? Are we is this like? Are we rolling this out?

E

Is it fully rolled out um that kind of stuff and um I'm kind of I'm, like personally I'm kind of like I, think it's a good idea to have a generic kind of way of having a link to um like further information, I think from a developer kind of usability perspective? That's super useful. The thing that I'm not sure about is where it fits into like open feature, because I'm not sure like at run time like home feature, is essentially like a runtime thing.

E

That's evaluating flags, and so, like a an engineer, the only place where an engineer is going to be kind of like see would get access to this URL would be in the logs or something I suppose or like in an Hotel. Spam is, is kind of like the obvious place, so I think I, guess I'd be I, don't know, I know what other people's thoughts are about, but, like I couldn't quite figure out where this would fit into um into like this, the spec or into the the SDK.

H

It seems like it's a concern for providers, not a concern for for open feature.

D

Yeah yeah I I would agree, especially after kind of Pete explained, because I took a different approach to it, but I think what Pete was saying makes more sense. However, we would, it would have to be kind of part of that contract, though between a provider and the SDK, um you know, so it would be some kind of like you know, provider metadata or something like that, um and possibly even having like kind of conventions around that. If we wanted to go that route so.

H

I'm thinking not as a open feature provider code thing, but like the the people who run the service that that manage the feature Flags. Actually that's where the URL should live at runtime I, don't think you should have it yours, you should have your target. You should have your feature flag key and then you go look up your feature flag, be in your source of record.

E

I, the the I I agree. The only thing that I, the only twist that I would add, is having like a defined kind of like semantic for like this is the issue URL that openfeature knows about and then can put into open to, like an open, Telemetry span um like by default would be really handy for, like you can. I can definitely imagine a lot of scenarios where a uh a devops engineer or someone who's debugging, something and they're looking in their observability platform.

E

They see this feature flag is here and they and having like a link right there in the in the span that takes them to the details of this flag would be extremely helpful. If you're, like your hair's on fire and you're, trying to decide, can I turn this flag off or not like I. Think, that's a really useful use case, and so having some Mech like having open feature, have some mechanism that will help kind of thread things from the provider through to like the observability platform.

E

The feature flag provider through the obserability platform, I think, would be super valuable um and yeah. Actually, that's.

D

That's probably a better approach instead of trying to have jira tickets and all kinds of other stuff into here, just say like if we could provide a link back to the flag in a particular tool, which is something that I had played around with it's not currently possible. Unless it's like a in some situations, it is because it's just like literally like the URL contains like the flag key or identifier and it'd be trivial to build the URL.

D

In other cases, it's using, like you know, goods and stuff like that that we wouldn't have in the SDK, so it doesn't mean we can't do it, but it might be challenging to do you know across all providers, but we could add support for it and then, as providers you know have the ability to you know include that information. We could certainly do it, which I think would be super helpful in like the details, because then you could use it in Hooks and Telemetry and dump it out to the yeah.

D

However, you want to use it at that point. I.

G

Think I think I do well. I would worry about this, potentially maybe more for client-side sdks from a security point of view, I I, don't I. Don't really think that it's the domain of I mean I I can see how that would be helpful right if you're, yeah, everything's, burning and you've got like a perfectly crafted link to exactly the piece of information you need.

G

But um we've had a lot of experience of of um you know people people do crazy things um uh and as I guess for server-side sdks, maybe it's not so bad, but for client-side ones. I would be like I wouldn't want that data going down the wire to like a browser like that just seems like a recipe for disaster.

E

I think that would be a like a a thing for the provider to decide right like the provider is the one that would be servicing it into kind of like open feature anyway, so that would be like ever ever. You know if you're building a yeah I mean it I, don't I, don't think I think having the option to do. That doesn't mean that it would be uh have to be followed.

E

I think I'm, trying to think through whether that there's some wrinkles there, where like if I, was making a generic like JavaScript provider I'd, have to kind of know in the provider whether I'm running in well, I, guess, I, wouldn't I would just have to you would never even put it on the wire right, so I think you just have the provider would just have to have enough flexibility to to say, like maybe I'm, going to get this information and, if I'm running on the server but I'm, not if I'm running on the client, and so, if I have it on Surface or something like that.

E

G

Mean it it is interesting like we like. We do have like a description field for flags that people can put things of that nature. We definitely don't send them down to the sdks like we don't send that you know that data down, um but I haven't said that it you know, I could see that being potentially useful yeah.

G

If you were, if you were running on a in a secure con runtime like on the server side, um so maybe there isn't yeah, maybe having some sort of I, don't know like yeah some sort of flexible yeah, either a blob of text or a set of KeyBank like a context that the provider could yeah push information into that that that might be helpful. But my worry there I would be worried about security but like like we, we have like two different types of API key.

G

Now we've got like one: that's meant to be public and one that's really not um so I guess yeah. If other providers have that Paradigm, then that that works better, but yeah I think you're, right, I think by default, you wouldn't want to do that and let the provider do it if they are really certain that it's not gonna, like you know, leak later that you don't want to leave. Yeah yeah yeah.

D

I was just going to suggest, if you don't mind, maybe open up an issue, maybe as as a research task like in the enhancement proposal, yeah and then link it. We can chat about it. There I mean I, think, there's I, think it's a good idea and probably to keep the scope at like how do we get back to like the flag management system for this particular flag, and then everything else, I would say, is probably out of scope, because I think you could get all that context from like the flag management system, because.

G

You do have the flag identifier as well right, like is kind of more of a convenience than anything yeah.

D

But like I mentioned not every system allows you to go directly to you.

E

Know because you need.

D

To know the environment, in some cases.

G

Maybe right, okay,.

H

D

Other complexities- and it would be an optional thing, of course, to possibly even something that's disabled by default. You could enable it no like there's a lot of ways. I think we could. We could introduce this feature, certainly in a non-breaking way too I.

E

Mean I the the way: I yeah I'll. Definitely I'll open, um like a research item, I, think the way that I kind of putting my kind of like spec writer hat on or spec uh whatever, um like I'm trying to think like I, think there's a way of doing it in a very kind of light generic, giving some kind of very light generic way for a provider to put metadata in that then gets surfaced into a hook.

E

That's then kind of available in a hook right because that's like the two, the two kind of like places where you you need to be able to make that connection, because if you can get the data from the provider over to the hook, so the hook can then push push it into open, Telemetry or write it to a log or whatever.

E

Then someone uh someone who's working at company X can use that that capability to wire it all up right, like you've, just basically provided the connection exactly so I think we could write, make something super super generic.

E

That's nothing to do with like issue like URLs or issues or anything just like, uh like Justin's kind of like mentioned, like just like a bag of like key values, and then um someone would write like a like a kind of collab like a kind of collaborating thing where, like the the provider dumps in like the issue, URL or the devops page or whatever, you know like the thing and and there's just like a contract between the hook and the pro and that specific hook and that's to be provided.

E

That knows like this key value is like the thing I need to to put this URL into the open, Telemetry span or whatever so I feel like that doing. That is like a super uh General thing and then, if it becomes a useful thing that lots of people are using for specific things, then you can start putting some semantics and saying like in this key value pair.

E

um This, like basically, the equivalent of like kind of like those kind of defined um attributes in open Telemetry, like this specific key value pair, is where the URL goes for the flag or the like. The issue. Url yeah.

D

Exactly and that's what I was going to say, it reminds me a lot of like the semantic convention in hotel or you can pass in whatever you want, but then they have a system or saying like this has specific meaning and it's meant to be used in this purpose and I think that that could work out well so yeah, if you don't mind uh creating that that research issue, maybe linking it in in that thread, and maybe more generally in the slack Channel yeah, and then we can just kind of do a brain dump of ideas, uh but it isn't I think it's a non.

D

You know breaking thing and something that we can kind of work on and certainly add. uh You know when, when we feel is appropriate so.

E

Which? um Which where's the right repo to put that.

D

um Probably the uh open feature like enhancement proposal, that's kind of been like where we've gone like when it comes to like spec changes and stuff, like that, it's in research.

C

I think it was.

D

Renamed to uh ofp uh enhancement or whatever uh too early I, guess I'll link it to you yeah. It is.

C

Okay, yeah perfect.

D

Cool um and then uh well uh anything else that you wanted to cover Pete.

E

No, that was it.

D

Perfect thanks for bringing that up um and then I guess once you do have that, maybe if you could just pop it in that uh channel, so Sean sees it, I think he's in Australia. So it's probably very unlikely. He'll he'll join any of these meetings, um so we can just collaborate with him. Async, yeah, well um David looks like you're next.

B

Yeah, all right so I'm bringing the fun topics no just kidding um so in the next yeah.

B

Finally, I'll have more time to dedicate also to to open feature again and I was hoping to help set up an outreach program and by that I mean even just collecting all the you know, speaking engagements collecting all the different uh slide decks and making sure that these that are, um it's a visible, whether it be through social media, whether it be even just having them in some way displayed on our on the doc site or on the blog post section, so really driving a bit more visibility and and and maybe even looking into how we can uh increase adoption, um obviously happy to work with.

B

Whoever is open to this. If you want to write a blog post uh or things like that, I'm I'm happy to work with you um and help out where I can um and yeah and then yeah I don't know. If there are any questions here. Doubts comments.

D

D

Because it's related, um we do have that issue. uh We keep opening every two weeks in the community. Repo I believe um where you can just add in you know, events that you'd like to have the open feature like Twitter account mention, um and so it's kind of a nice way to to have those scheduled and then then release. So, if you you know, maybe Thomas wants your blog is merged in. We could add that there, for example, if a provider's developed anything like that um would be nice or speaking engagement.

D

So just keep that in mind. um It's pinned at the top of that that repo, so you can go ahead and just add a little uh little comment. There thanks yeah yeah.

B

And that's. This is also to give visibility, as you just said, to to all the providers in there to make sure that you know everyone knows who is a part of the project and and can uh and that it can be easily found um then yeah. The other topic is the bootstrap governance I'm, just bringing this up because I'm, that's I'm, a program manager I like to annoy people.

B

So that's that's kind of in My, DNA, um I I know Alex uh Jones put in his uh sorry put in the candidacy um yeah and looking forward to see. If anybody else wants to volunteer other than that I think, then it would be nice to whoever is present. Then at kubecon again have a discussion about this in person, and we can make sure then to transfer this all over in some async form. That way, everyone else can also be a part of the conversation.

B

That's kind of it on my side, I, don't know if you have any questions again, let me know uh also be trying to help uh Mike a bit more with some. Let's say: infrastructure, basic enablement, things um yeah.

D

So perfect, thanks David uh looks like Alex joined, so we just have to briefly recap: everything.

G

D

No, no worries: um do you have any topics Alex.

F

um I I guess one or two just just really like kind of boring logistical things, so I think uh oh, like or somebody else mentioned, having some open feature slide deck that'd be kind of cool. You know, like a reusable thing: I've got a whole bunch of things, I'm going to talk about and I'd like to have it looking like everyone else's.

F

um What else? No I think I think we're making good Headway I'm excited for cubicon, and we could talk about marketing actually a little bit that could be kind of fun. I want to put a QR code in my talk so that can drive people to the site. I thought you know we could even look at the doc section.

F

I was speaking also to people like James and others around after kubecon I would expect to see a little Spike on activity for like open feature operator and and flag D and the sdks client libraries just be really nice to make sure that before that, we've just gone through them and just double, checked and made sure that all the readme's look good, and maybe we can do kind of like a code. Freeze, if that makes sense, you know what I mean like in the next couple weeks.

F

We just say we put down the tools and we just look at the docs and make sure that everything feels welcoming and people would be easily guided to the right stuff.

D

Yeah I think that makes a lot of sense um and I think we're in a good spot. For that too. uh You know now that we have like those those last breaking changes. That I was aware of, at least from the sdks, the ones that I feel most confident with that.

D

That I think will be ready by kubecon or you know, java.net go and JavaScript, so I think those are the ones we'd want to focus on and and have as many providers you know up to date and ready to go by then um yeah, like you mentioned, docs, are extremely important to make sure that onboarding experience makes sense.

D

The initial Impressions all make sense so yeah, if, uh if people have time to kind of review that the happy pass and if you have feedback you know, certainly let us know and now's a good time to do it um and then uh for a QR code. I was also working in my slide deck as well. I I, don't know if, if it's uh yeah, maybe we just want to create like a QR code that we can share.

D

Potentially um one of the tools that I found has kind of like kind of pretty looking QR codes, and it also does like uh scan tracking, which may or may not be attractive but uh kind of kind of an interesting potential option there as well, but it's kind of cool because you can have like the logo in there and it's kind of more bubbly.

D

Looking QR codes and stuff like that, so A couple options that we could share and then maybe we could, you know, uh make available to anyone that wants to use that type of stuff.

D

Was there anything else that I missed I mean I, think that that's all good I think that you know I'm totally on board.

G

One question um uh in terms of like tracking uh analytics Pages, the docs um I know some sdks languages, it's easier to track activity downloads and things like that than others like Java. It's not really possible as far as I'm aware like is, that is that fun and like is there like some sort of analytics tracking on the documentation side, for example,.

D

So we do not have Google analytics or anything set up, although we probably could- and maybe should it's something that, like even the cncf would like access to something I've been you know not reluctant to do is just one of those things like I only want to do it if we're going to use the data, um and so I I mean if, if there's demand for sure I think we can go ahead and add that stuff to the docs.

D

So it's not currently there, though, and in terms of downloads of the sdks I mean we haven't, really talked about it I mean I would for, like you know, npm, for example, you see downloads. You know same thing with, like uh you know, containers in the GitHub container registry things like that. um That would be the only way to track it at the moment.

G

Yeah yeah, so Google, sometimes in this sort of context, sometimes feels almost a bit grubby, but it's just if it. If there's, um if there's like energy being spent trying to get people to you, know the GitHub repositories, the documentation and stuff they might. If you know, if you don't have it you're, never gonna you're, never gonna have that data right like so.

G

There might be things in there in terms of I, don't know where people are coming from or referrers or things like that, like you know, I I, every every couple of weeks, I look at um larger, GitHub repos and see um whether github's so frustrating because you they show you that data and you can't track it and then it goes so you have to do it regularly to see like sometimes we get I, don't know someone's written about us on a Reddit post that we didn't know about, and we only found out about it because uh it was an accurate Hub.

G

Reference kind of thing.

C

Yeah there's a couple: there's a couple things that I I do. One is um my Google alerts set up personally, which, if you're really interested in the project you can set up and it's kind of nice, because you'll get alerted for even like a big Hacker News article, for instance, if it gets popular.

C

So that's not a bad thing to do on a personal level um as far as uh downloads, um ossrh does actually work that for the Java, um so you don't publish directly to Maven Central, but it's kind of annoying, but you publish to a smaller repo that mirrors into Maven Central, not job and and that smaller repo is, does get information pushed to it or or that or it kind of has some way to track the downloads.

C

So it's just not public, um but we can use a credential, a shared credential to actually look at our download statistics in Java. It's public and npm I, don't know about go, but it might not be a bad idea to kind of centralize these this data somewhere. Yes,.

G

I wasn't aware about that with Java. That's interesting, um but yeah I mean I, know like it's been useful for us like just the npm traffic and stuff. That's that's been useful to to see um as a vanity exercise. If anything else, you know.

D

Yeah, the other thing that that this reminded me of is um currently when, when we're registered on that cncf landscape, page I had to point to a single repo I chose the community repo, um which you know because of just the state of it right now. It doesn't have a ton of activity from commits or necessarily Stars. So it's one of those things so I think we should try to add some more activity in there where possible.

D

Without you know forcing it, but also if you haven't, if you don't mind like you, know starring that repo, because that's kind of the one that people see when they go, look at it on the cncf landscape, so um yeah, just just a quick aside um that that's that could be a a bit concerning. If you go there and you see like the the commit activity kind of drop off, because we haven't been super active in that repo and obviously like we haven't, pushed people towards like starring.

D

It necessarily which is you know, one of the metrics I guess uh in GitHub at least so it's it's. uh It would be handy if we could keep that in mind. As we're working.

D

Perfect, um it looks like that's everything on the list. uh Is anyone else have any last minute topics perfect all right? Well, the next couple weeks will be pretty exciting. I would say so a lot going on so uh yeah just.

F

Forgot, disregarding that quickly, Michael have we organized meet up as a team.

D

I meet up in in Detroit.

F

Yeah, like maybe we could just put some, you know, ideas for a time and a place at qcon and do a little unofficial office hours or something.

D

Yeah I have a sign up form too so I know you're gonna be there, so you can. You can sign up if you want or don't worry about it, because I know you're going to be there, um but we do have a few options. We also have the booth um that will be. You know that's in the middle of the day, um there's also that that uh Cube cruise that we can meet up at and then um certainly it's probably a good idea to you know.

D

If people are available and willing we could we can meet up either at the event space or even at the diners Trace office, because it's within walking distance of the event.

C

Did the dinosaur's office to be clear.

D

What did I say? No I.

C

Thought you just said office, but oh no.

D

Yeah cool um but yeah I think that that's a good idea. We should definitely use the opportunity to meet up. um There are a few things, I think that are already scheduled, but um you know any dedicated session time, certainly more than happy to to meet up or try to coordinate. So just look for Todd's silver hair. That's perfect!.

F

It was actually kind of cool we do at the Donna Trace uh offices, because then we can include people who aren't at the event, so I think that would be nice yeah.

D

That's true actually um so I can work on that. It should be easy to get. You know approval and it is kind of cool because the the office is in uh an NFL stadium. So it's kind of a a unique experience to to walk in through, like the main gate, and you can see the field and all that stuff and you head up to the office. So um the office itself is not that exciting, but the uh experience of getting in is kind of cool, so awesome all right anything else.

D

All right, guys, yeah appreciate the time have a nice day, I.

H

G

Thanks guys see ya all right.