Hey guys. I’m willing to stick my neck out for the community. My neck twisted when Kadira decided to stop, but its open source continuation triggered me into a mindset. What if I would host the open source containers and mongo database and made it a SaaS again. Companies might be willing to pay for additional development, but I’m looking into how I can shape it in such a way to strictly stick to the open source ones. Whenever a company pays, they pay me for my time, but also contribute directly to the community. They would literally pay a guy to contribute to the community to have their request be prioritized.
Its right now still an early thought. What do you guys think and what would be a good approach that suits us all? Maybe extend the idea to make a pool of developers with some new kind of payment / contribution model?
Some brain-farts:
Development contributions to features requested by companies will result in a payment?
Keeping it strictly open source. No such thing as enterprise grade Kadira versions etc etc. Available to one is available to all
I think that would make a lot of people very happy. How would you pay for the servers and dbs to keep it up and running though? You’d need some sort of monthly subscription from users, right?
Yea, but I do spot a need there. I’m looking into how I would approach this. Is it acceptable to have a small profit from it because of the money spend and risks involved and into what degree do I need to provide an SLA. If I need to provide support and/or an SLA would it be worth it, etc. etc
If there was an easy way to tie costs to directly resource use, that would be ideal. Given the nature of Kadira, that seems pretty do-able. And I don’t think anyone would object to you building a profit margin into the subscription cost, given the time you’re saving them by offering Kadira as a service.
@sashko Said himself that he’d be happy if someone did this.
It certainly seems fair to take out a profit for your efforts in administering the servers
Maybe a bit less than Kadira charged as you didn’t create the software
I was paying for Kadira a while back (they must have stopped charging around the time they said that they were going to discontinue the service) and I’d be more than happy to pay someone to host it as a service again.
We are a small but growing startup that uses Meteor.
We would gladly 1) pay a monthly fee, 2) pay for hourly development.
A start could be to pay you to get Kadira setup to work in our use case. You would be free to document the entire process. Whatever features we would require, we would pay for on an hourly basis, and it would of course be open to the community.
If the hosting cost is a risk to you, we might help absorb that too if we agree to some arrangement.
I definitely agree that there is a market for Kadira to keep going as a SaaS!
… and I definitely also agree that many (including myself) would pay for it!
If I were you I would NOT offer a free tier as kadira did, but maybe go for a very cheap starter package for smaller apps.
My guess is that many Meteor apps out there are pretty small in size but still need something like Kadira, so if you charge say 7-10 bucks per small app my guess is that you can definitely cover the costs of hosting and small parts of your development time.
I love the idea! Kadira Premium was also in my planned expenses… You could see how Gitlab work out their expenses/revenue as it seems to work very well! I think their way to handle it is very good and I support them for that.
Basically, they are open source free + closed source premium. They host a limited free version, you can pay to have bigger/better options, or you can host it yourself. They also offer a licensed version with more options suited for big companies…
I don’t think having a free option is a must though, if you’re not an ambitious company looking to grab market share. Maybe a bottom tier of 3 bucks / month, where you can only register 1 app, with 3 day data retention or something.
And then the other tiers being similar to Kadira, but you could cut their prices since they wouldn’t need to support the free users or pay for the creation of Kadira.
I will take a look at the Gitlab stuff. Right now, I’m focused on figuring out what the best way is to deal with payments and I need to do some risk assessments. If that’s sorted out I can put it online. I cannot guarantee any deadlines unfortunatly, because it all depends on my spare time availability.
I get your point about the free tier, but I do think it is needed to bring some mass to the project. Having 3 customers pay 100 bucks per month is more risky than 100 customers paying 3 bucks. Ok, the loads are the not same so the running costs too but it could be depending on different limits.
My point is that with the free tier, you are most likely to see people use your paid solution. If I were to spend time to setup kadira myself, as I had to do with a couple of services just so I can lower costs during development before I can sell my product, chances are that I will not want to pay for a third party, as I have already all my setup, and I know I won’t disappear just like Kadira did… Having a free tier means that I could reliably setup for development and when I ready to go full speed I just have to turn the key.
The different expenses coming from free tier could be covered with some (small) advertisements and the likes. Also that makes more “forgiving testers” available…
I just want to leave a message here. I’m still working on making it all happen. Part of this is getting familiar with the codebase and creating a plan with actions. I would like to share some ideas right now:
Costs of hosting it
Like all production apps, I need to determine the costs of hosting the application plus the mongo database. I’ve a rough idea of what it takes and have some samples running on AWS. Ofcourse I need to see how it performs when multiple applications are connecting and putting load on the system. With that in mind I can calculate some of the pricing models.
The thought right now is to have a free tier available, but a bit different then before. It might involve unrestricted registration of an email address and maybe the creation of 1 or more apps But with a slightly more restricted timerange then it is now.
I’ve had the personal experience of searching for monitoring dashboards. Kadira shines compared to other dashboards, but I might have skipped looking at it in the first place if it didn’t have a free tier.
Code base
Like in all applications, there’s stuff you like and don’t like. I’ve seen some things that I like, but also things that aren’t that trivial to change, maintain and scale easily. Its for now not really a blocker. I just need to get more familiar with the business rules and I might want to create some tests.
Data pipelines
One thing that might become a bottleneck in this dashboard is the way it handles metrics. Its on a set interval which causes spikes on CPU usage. I’ve always built dashboards like this using a stream based approach. Again this might not become an issue depending on number of apps, etc etc etc.
Its going to be a spare time project
An important aspect to remember is that whatever hosting I can provide will likely be on a spare-time project basis. I’m right now working full time for another company and can’t respond immediately whenever something fails. This needs to be clear and will likely have impact on the pricing, because although it will be a SaaS, that last S will not always be 100% guaranteed. Or at least not within a timeframe of lets say 4 / 8 hours.
I think that the strongest open source related or backed projects are the ones that act on transparency. That’s also the reason why I’m sharing this. The main goal of building this SaaS is gaining experience making my name visible in the community and providing some value to the community. Any thoughts, suggestions and experiences are more then welcome.
One thing I would think about is allowing for it to be an “on demand” service. My experience with Kadira has been that I found much more value in it as a diagnostic tool than as an monitoring/alert tool.
I’m sure this isn’t the case for every project, but many times, I didn’t need it running 100% of the time, and I feel that most devs with a working knowledge of Meteor/PubSub/Mongo will already have an idea of what could potentially be bottlenecking their app and not be combing through weeks of data looking for patterns.
If you anticipate a sizable cost dedicated to what are essentially “idling” apps, maybe something to consider?
Contact me I will take care of the hosting and the platform (I already have made a platform alternative to embedly called EmbedAPI.com) accepting payments Stripe,Paypal and Bitcoin.
I already have 70% of this done so if you want partnership on this we can make it faster since I have never used Kadira.
Sounds nice! Can you elaborate on your plans? PM if needed? I mean, if we can align our roadmaps / ideas and vision on this project that would be a good starter.
