Small thought here:
If we spend 150k on 3 devs to get us WASMER we will never have these scale issues on API nodes again.
All i heard in the past months is that nodes are so busy keeping up with the actual network that there is little power left to serve queries.
I wont vote yes on a proposal that doesnt come even to close to solving the problem and instead tries to stem the bleeding.
What we need is the WASM engine upgrade and the new Gen Intel CPUs to be supported, on that i would spend 150k.
Let’s all take a step back, and continue this discussion in a productive and friendly manner.
I understand you believe this is a priority of the highest order @stevenchang5000. I’m not sure though that starting the title with ‘Emergency’ is the most productive approach. Over the past year we’ve seen the creation of a decent number of new api clusters that dApps can use (although not for free). If I understand correctly @anon60841010 has recently opened up the Spartan API after the node count was further increased, so developers should even be able to get free API access there.
Keplr has been having problems in the last weeks, luckily it now has a feature to enter different api endpoints so that should solve some of the issues. We still should aim for an optimal experience on the default cluster of course.
Let’s use this thread to ideate on the improvements we would like to see and how we can get there in a friendly, respectful and constructive way.
Money saved is better than money badly spent
This conversation has been very chaotic, but I find myself siding with those who say it costs too much.
While I do agree that some development needs to be done, this will probably cost 20k at most. This would be to develop a website tracking the status of the API, handle monitoring and restarting nodes, and maybe showing usage over time.
What are the remaining 160k used for? Secret nodes was able to buy 17 nodes for 80k (which I completely believe given the price of the hardware), then setup load balancing etc.
I haven’t seen how spartan is setup but I assume that Ian and his team set something up with chef or ansible and it takes almost no time to spin up, manage and keep a new node in sync with the others once the hardware is assembled and the base OS is installed.
Even if all of that were to be developed, and you would spend 20k on collocation costs (which you would not) there would still be an additional 60k that’s unaccounted for, especially if it’s recurring.
Button’s API and site will be a gateway for api providers.
Keplr =>
Button’s API =>
=> Spartan’s public API
=> Secret Saturn
=> Agora
=> Ronin
etc etc.
Our API will distribute the load to the different providers, remove ones that are temporarily down and re add when back up, keep track of usage of each provider etc.
This won’t fix it, it’s really about server overload and management and decentralizing points of failure.
Honestly why not? The network is in shambles, we are bleeding users and developers. Just wait it out?
Pleaseeeee, anybody, take that 20k offer and get this fixed. Been waiting for a year for this. I’m fine if somebody else does it. But if nobody will, we need to get this fixed.
I don’t even want to do this. I want to work on dapps. Only puttting this out there because it’s not getting done and it’s at a point where heaps of people don’t want to deal with it anymore.
Hello,
I’m writing as a long-time user of SCRT network, a BUTT holder, and part-time agent who worked with Brendan a bit in the past. I’m not a developer, although I am an artistic type outside of crypto.
Opening my Keplr wallet today, I see red exclamation points popping up near some of the assets. When I reload it, it works better. However this still feels like an unreliable beta mode which would put off a lot of first-time users who expect a smoother experience (they might not even try reloading the window.)
I’m aware API’s are the issue and have been for some time. I’m excited about the mobile Fina wallet, and am wondering if there could be a desktop web-extension version of Fina, but Steven says it’s the servers, not the wallet, which is the issue. Moonstash says there are other API’s, but poking around in Keplr, even as a long-time user, it’s not obvious as to how to change the API. Is there a way this can be adjusted automatically by Keplr according to which is working best?
I’m guessing “endpoints” under “settings” has something to do with it, but it’s not obvious and comes with a warning “experimental feature.”
A note about personalities - I believe in Steven as a developer and as an artist myself, I know creative types tend to come with dramatic character. I hope we can learn to be tolerant and patient with each other especially given the frustration coming from the current bear market.
I have no idea what’s appropriate when it comes to funding development. However, I’ll leave you with this quote from an old Andre Cronje blog which impressed me: “I know traders and investors have a saying ‘do you want to be rich, or do you want to be right.’ I never understood this saying, but I finally do. Unfortunately, I want to be right… Bull markets, cause bear development. And bear markets, cause bull development.” - April 2021
I recommend reading the rest:
source: https://andrecronje.medium.com/bull-market-bear-development-b6ede96d4c82
Well…there have been a few opportunities that you wanted funded that SLABs turned down that could be good pool candidates. It doesn’t seem like the place to talk about it necessarily, but to me this one just sounds very expensive for things that we’re supposed to already have multiple solutions to.
There is no solution to it right now. If there is one, please apply it. That is the point of the proposal.
There is nothing worst on secret network than Keplr API issues on every single launch product, be it an app or an NFT mint on SN. And it goes beyond that, when the whole market volatility goes a bit wild Keplr breaks the first of them all and there is nothing a regular user can do.
However, people with technical expertise are able to get around Keplr API downtime and send their TX directly on their own nodes and such. They have so much to gain when the majority of the users (including Binance) are on technical pause.
Playing devils advocate, those gains are:
Generous arbitrage opportunities
Mint 3 anons on multiple wallets
Validators collecting generous failed TXs
At the beginning was more of an indicator of a positive growing demand for SN, and at the time I was so naive to believe that there will be an immediate investment to deal with that bottleneck.
I am not the one writing the cheques but my fantasy based estimation tells me that millions of $$$ were deflected from entering into SN and other millions of $$$ are exiting SN because such a bad Keplr experience. $180k is a drop in the ocean to give you some perspective. Moreover, the issue was not dealt with sub $1 SCRT, nor at $4 SCRT, neither at $10 SCRT, and back to sub $1 SCRT and we are like… yea man at least there is a 10M SCRT Tarantino NFT hidden somewhere.
I support button group for a long time, some of the few left living candles on secret network with a genuine altruistic hearth for the network.
You all would have approved any solution for this issue for $1M or even more during the two year bull-run we just had.
I’ll repeat what I said to Ian even if it costs me another 3/4 delegations.
You cannot have an endpoint that satisfies all users.
He has since learnt that, and I may not be able to convince you without you also trying.
So instead, please provide a breakdown of how you came to that number.
For the 3 months development, please breakdown the tasks and expenses involved. Estimates are fine.
I too support your efforts Steve, and hope you will continue to build cool stuff, I’m just not convinced this is the panacea.