Meta: future of Rusoto
matthewkmayer opened this issue · comments
Effective immediately I'm capping my time spent on Rusoto to one hour a week. As of February 1st, 2020, I'll only be available to help with hand-off items.
It's been a fun and challenging ride, but it's time for me to move on. See my blog post for more info.
Call to action
Rusoto could use more maintainers! Look at #1496 for more information on what's expected of maintainers. 👍
Thank you, contributors, maintainers and users. I look forward to seeing what happens next! 🎉
@matthewkmayer Thank you for all your effort!
Is there a list of current maintainers anywhere? I'm wondering, if someone implements an important feature and the merge request is forgotten for some time, who can they poke?
On a somewhat related note, I don't think there's any chat for this project? What do you think about setting up a chat, perhaps on gitter, where it would be possible to find someone to look at/discuss proposed changes? Don't know about gitter, but I believe discord allows to create e.g. a @maintainers role. Maybe something like this could be useful?
I’ve considered setting up a Discord (similarly to other Rust communities, including the Rust project itself).
If nobody beats me to it, when I get time at office today I’ll post a list of maintainers here.
@iliana Did you end up setting up a project Discord? There's a bunch of stuff around Rusoto that I'm happy to contribute to (accepting that it's not my day job either!), but likely wants broader coordination and discussion, including helping land the switch to Tokio 0.2 and std::future::Future, and cross-platform CI and testing, perhaps using GitHub Actions.
Haven't yet, that's on my list of things to do next week.
Discord set up: https://discordapp.com/invite/WMJ4DWp
Will get it in the README.md (unless someone beats me to it!) next week :)
Here is a blog post I wrote about the direction I'd like to gently steer this project toward: https://linuxwit.ch/blog/2020/02/the-future-of-rusoto/
Comments welcome here.
So... Is this project still alive? :)
This probably a naive question, but is there something that we could do as a community so that AWS officially endorses the maintenance of Rusoto?
The experience from my own projects is that without official funding it is hard to keep the momentum going with projects like these. My own employer, although being overtly open source friendly, is under pressure to deliver products and so has little time for infrastructure development. I do get a lot of help from academics, who can generate papers from open source projects and especially PHD students who will benefit form the experience.
It seems that AWS do fund Sean McArthur and other long standing rustacians and so the future may not be so bleak.
I'll see if I can get in touch with one of the AWS rustacians to see if there is a possibility of funding, but don't hold your breath! I would suggest joining Open Collective or similar to provide a funding umbrella to buy some time for Rusoto developers.
This is purely speculation but based on recent openness about aws' investment in Rust in would not be surprised to here something at this year's reinvent that might help answer this question.
What's the situation with even basic maintenance like upgrading dependencies? That's something that even people who don't know much about AWS but who use this crate (e.g. me) could help with.
One can use their own fork of Rusoto to get changes such as upgrading dependencies.
One can use their own fork of Rusoto to get changes such as upgrading dependencies.
I've done just that here: https://github.com/MaterializeInc/rusoto.
So far, I've bumped all the dependencies to their latest versions, with the notable exception of Tokio. I'm in the process of releasing these changes as new crates named with an "mz" prefix (e.g., mz_rusoto_core). Unfortunately this process is going to take several days unless the crates.io team is willing to bump up my crate publish rate limit temporarily.
I've sent you an email as well, @matthewkmayer, since I'd be happy to instead help out with shipping a new official version of Rusoto.
Thanks to @iliana and @matthewkmayer I'll actually be helping out around here instead! I'm not able to do much beyond keeping the lights on—reviewing dependency bump PRs and obvious bugfixes—but hopefully that is enough until a more sustainable maintenance situation can be arranged.
v0.46.0 is poised to ship in the next few days. The last piece is the upgrade to Tokio v1.0 in #1880.
I'm going to try to keep to a monthly release cadence from here on out, but as stated previously, I'm limiting my attention to obvious bugfixes and dependency bump PRs. Here's the tracking issue: #1898
We wish you the best. There is a definite need for this project.
@matthewkmayer - I saw the thread on #1496 which looks like you found replacements? Are you still looking for maintainers?
Hello guys, how are you?
I'm studying Rust lang and I want to try help with Rusoto.
My knowledge with Rust is limited but I have experience with NodeJs, PHP, Linux servers, aws, etc..
We have any product backlog with priorities? What is better way for me get tasks?
Thanks!
AWS rolled out an ALPHA release today for their official rust sdk. It doesn't have all the features of rusoto yet but it does include standard retry logic and they have a roadmap up
https://github.com/awslabs/aws-sdk-rust
https://aws.amazon.com/blogs/developer/a-new-aws-sdk-for-rust-alpha-launch/
AWS rolled out an ALPHA release today for their official rust sdk. It doesn't have all the features of rusoto yet but it does include standard retry logic and they have a roadmap up
https://github.com/awslabs/aws-sdk-rust
https://aws.amazon.com/blogs/developer/a-new-aws-sdk-for-rust-alpha-launch/
Guess who wrote that blog ! @iliana of Rusoto fame did ! So, there is some hope that the alpha version is still pretty usable.
Well, I lied about the monthly release cadence—learning that AWS was releasing their own SDK was rather demotivating, sorry! But the good news is that v0.47.0 will drop later tonight with some dependency bumps.
As mentioned above, an official AWS SDK for Rust is in alpha! I don't plan to keep maintaining Rusoto once the official SDK is stable. But I'll plan to swing back through later this year and do another round of dependency bumps.
Regardless of the official SDK, I think that I can say that we all appreciate your efforts very much.
This project has got the ball rolling and made a big difference to our projects.
Please direct your gratitude toward @matthewkmayer, @iliana, and the numerous other contributors and maintainers over the years! I’m just doing the bare minimum here.
Really cannot even begin to express how valuable all of the work on this project has been. I can't imagine trying to integrate with AWS up to this point without it.
https://github.com/awslabs/aws-sdk-rust/tree/main/sdk
Aws sdk for rust s pretty complete now
@peterbean410 Yes, but it still has the following disclaimer in the Readme:
Please Note: The SDK is currently released as a developer preview and is intended strictly for feedback purposes only. Do not use this SDK for production workloads.
I agree that a "totally unofficial" library isn't a better choice, but in my case I already have some code that uses rusoto running in the production. Therefore I'm going to migrate to the official SDK only when it is officially declared as ready or if I encounter some bugs in rusoto.
Are you happily using the official SDK and recommend migrating to it?
@peterbean410 Yes, but it still has the following disclaimer in the Readme:
Please Note: The SDK is currently released as a developer preview and is intended strictly for feedback purposes only. Do not use this SDK for production workloads.
I agree that a "totally unofficial" library isn't a better choice, but in my case I already have some code that uses rusoto running in the production. Therefore I'm going to migrate to the official SDK only when it is officially declared as ready or if I encounter some bugs in rusoto.
Are you happily using the official SDK and recommend migrating to it?
I also believe that if you already have rusoto in your code, it makes sense to stick to it for now. In my case, I m working on a new project and it doesn't have this dependency. For my usecase, the preview version of aws sdk just fits my needs and I m happy to just lock the aws-sdk version to v0.6.0. Of course, we surely need migration when if comes to official releases, but at worst, it takes the same effort as rusoto. Optimistically, it takes less effort than rusoto as there could be fewer breaking changes.
Are you happily using the official SDK and recommend migrating to it?
We have switched to the official AWS SDK at @MaterializeInc and can enthusiastically recommend it. I am no longer maintaining Rusoto, and neither is anyone else, it seems.
Thanks for your opinions! I will probably wait for one more release and then also switch to the official SDK. 🙃
I'm currently releasing v0.48.0, with a dep update for hyper-rustls and adding ap-southeast-3.
It is extremely likely that v0.48.0 will be the final release of Rusoto. If you're waiting for a sign to move to aws-sdk-rust, here it is.
@iliana Thank you so much for all your work on Rusoto. (And a huge thanks to everyone else who has worked on it as well.)
Would you consider merging #1979 into a point release? The hyper-rustls version upgrade PR broke the rustls-webpki feature, and I'd love to have that feature working in the final release.
I think that’d be reasonable if it’s just core that needs a new release. I’ll look tomorrow.
@iliana I can remove the additional change that would require updating the other crates, to simplify the update. Thank you.