Does anybody uses Realm Cloud for production?

cloud

#16

@MartinHennrich

Thanks for the Firebase link. I’ve been working with that project since it’s inception - Firebase officially doesn’t support macOS and while this is an awesome open source, third party project, it would have a long way to go before it could be used in a full blown macOS production environment. It’s also a bit challenging to get up and running compared to the official iOS version

Keep in mind that macOS and tvOS are not officially supported by Firebase, and this repository is actively developed primarily for iOS

I would still like to hear from people that are using Realm Cloud in a production environment, whether is be a $30 plan or a $3000 plan. What is their experience? Cloud stability? Downtime vs uptime.


#17

@wildfisher when you have outages, what happens then? Does this just mean that synchronization is paused or does this mean that your users may have lost data?


#18

@wildfisher I understand your concerns. We don’t make support guarantees for the $30 users because we need to ensure that the enterprise customers’ problems are resolved as soon as possible. That being said, as you have noticed, we do our best to reply to all support inquiries, even if it takes a little longer.

Regarding cloud stability - we are doing our best to ensure all instances, regardless of the plan they’re on, have a “good” uptime. Obviously, we need to prioritize our resources and, admittedly, we will not escalate instance unavailable alerts that fire during the weekend for non-dedicated instances. That being said, outages as long as yours are extremely rare and historically, the combined uptime of all instances is more than 99.9%.

The cause of your outage was a sneaky bug in one of our native modules that manifested itself in rare cases when deleting users. We have identified the root cause and are working on a fix that will be rolled out as soon as possible. While this is not directly related to your question, I wanted to point out that we’re continuously improving the system and, while we don’t make uptime guarantees for the $30 plan, we do care deeply about each instance’s availability and have internal targets that we’re aiming for.

Ultimately our goal is to allow production apps to be hosted on any plan and many are. I understand and am sorry that your experience hasn’t been as great as we’d like it to be, but we’re absolutely focused on fixing the bug you hit and hope that you’ll have a much smoother experience once we do.


#19

@Sipe
Users do not lose their data. Synchronization is paused but main problem is that new users can’t register in the application since during registration they should be registered in Realm sever and create their database shared with colleagues.
Imagine the user downloaded the application from Google Play and can’t start to work with it properly.

@nirinchev
I understand the point about the inability to provide high-end service for $30 but it seems to me that another Shared Cloud $250 plan does not guarantee much more than $30 plan.
$250 plan also has 5x9 support, right? Low chances for 24-hours uptime.

The current bug with deleting users may be understandable, but what if the similar serious bug appears in the future?
For now, the main advice from support is ‘to not delete users unless absolutely necessary’
What is such function is included in the production application? Certainly, each user may require an ability to delete all his data.

I appreciate every help I get from support. Just saying that the current Shared Cloud is not suitable for production and Dedicated Cloud is too costly for small teams. It is costly not only for small teams but for most application targeted to mass market and charging users less than 20 dollars per year.
Also, Montly Active Devices pricing won’t work for mass market since 80% of users install, test and uninstall the application and all of them are counted towards Miontly Active Devices.


#20

@wildfisher thanks for letting me know. I do agree that pricing is a problem. It basically rules out using Realm Cloud for storing user data in freemium apps. Combined with that it is hard to predict pricing (no public info, no way to measure storage) it gets tough.

What I am building for my app is only to let paid users store their user data on the cloud. Free users will still be logged in, but can only use one device at a time and will have a simpler backup solution based on FirebaseStorage. For me it may be a way forward, even though it is a pain to have to build dual backup solutions because of pricing. I can still make use of some features of realm cloud for free users though, but pricing really gets in the way.


#21

Interesting this topic has been posted for roughly two weeks and there have been no responses from anyone using Realm Cloud for production.

My question is; What does that mean? Companies using Realm Cloud in production do not visit these forums? Or maybe they just don’t post responses? I have looked around and have not found any case studies or client testimonials re: Realm Cloud.

While there are compelling client stories, they seem to be focused on offline (first) solutions and give the impression these are self-hosted deployment or just proof-of-concept cases. They are also 2-3 years old.

Not trying to incite a riot here but are there any case studies, testimonials or other data showing how people are using Realm Cloud in a production environment?

Note that we are fully invested in Realm and have two parallel projects in the works. Once we switched to a Full-Sync Realm Cloud it’s been pretty smooth. (We couldn’t get Query Sync to work reliably and at the time, the documentation was inaccurate and outdated so we may give it another look)


#22

Transparent price tiers should be #1 on the new managements realm agenda.The “call us for a price” instills distrust and lack of confidence.

I’m using Realm Cloud in a very small production full sync setup. Currently all data in the realms can be recreated by a legacy system sync. I’m taking it slow to monitor technical issues and to get a feel for what the 30 usd Realm Cloud subscription can handle.

My plan B for a exploding cost and technical trouble is simply going old school again: sql db <=>.net core with grpc <=> clients with realm and own grpc sync.

I would love to go full throttle, but the lack of pricing transparency and a few strange events has left me very cautious.


#23

Hi Jay…

We did have nirinchev respond to this thread earlier (who is at Realm).
And yes: we have (even some rather larger) customers using Realm in production. In some cases we moved them to a dedicated cluster (vs. multi-tenant as would the $30 be).

I think your approach to monitor and gauge usage for your application is valid. We have seen a wide range of usage, and it is very sensitive to size of objects and number of writes and implementation details of Realm in the app. That is part why the $30 tier is rather generous.

Having said that don’t be shy and reach our to [email protected] to get a estimate for pricing for your project.

Pricing transparency (@void) will improve as well as we integrate more with MongoDB, specifically Stich and Atlas. There is work to do, though we are quite excited as to what is coming.


#24

@roberhofer
I am afraid you didn’t get all details of me and other people’s problem. I have contacted sales

The details why it is inappropriate model for most developers targeting mass-market and not ‘in-house’ apps are in this post https://forums.realm.io/t/does-anybody-uses-realm-cloud-for-production/2682/15?u=wildfisher


#25

@wildfisher do you need the realtime sync? I mean, I can understand the pricing structure per active device if you have a relatively small amount of users syncing high volumes of data in realtime. But I agree that it is expensive and basically rules out using it for any consumer/freemium apps.

For the apps that don’t need the realtime sync, I suspect that Google Firestore (and possibly MongoDB Stitch?) would be a better option. You don’t get the realtime sync, but if all you need is storing user documents and periodic sync for a consumer app (low bandwith, high number of users, potentially large data set that doesn’t change much), then I don’t see how this can be possible with Realm Platform.

I really hope this is where MongoDB Stich will come in…


#26

Perhaps I misunderstand the comment but the whole point of Firestore and the Firebase Realtime Database is live, realtime sync’ing so you definitely get that with Firebase products.

Firebase also provides a completely free development tier so you can spend time developing and not paying for it while you do.

My current .02

/beginRant

Has anyone tried MongoDB Stich? As is, what’s provided on Github flat out doesn’t work - it won’t compile at all and contains severely outdated API and examples. It’s built for Swift 3 so anyone that has XCode 10.2 can’t even try to compile it as it’s too old. I opened an issue two weeks ago and it’s not been touched (sound familiar)?

I think the bottom line here is the current Realm pricing model is undefined and confusing. There doesn’t appear to be anyone using Realm Cloud for production other than for in-house speciality solutions. The $30 tier is silly to charge a developer during development.

The path to MongoDB has not been laid out so it’s ‘uncomfortable’ developing with Realm at the moment as the rug could be pulled out from under us at any time ( anyone recall the Facebook/Parse deal?)

I would like to see the following changes

  1. Free developer tier
  2. A clealy defined pricing model. If you need to know how to do it right, look at Firebase Pricing as while you may or may not agree with their structure, at least you know what it is
  3. Clear path to MongoDB integration - at least give us the heads up on what kinds of changes we should be prepared for.
  4. Consistently updated Realm Documentation - especially around Full and Partially Sync’d realms
  5. More interaction with Realm both here and on StackOverflow
  6. MongoDB’s lack of attention to their OWN tutorials is very concerning. They should work.

/endRant

Note that we are continuing development with Realm as we feel it may be the right back-end product for our app solution but feel the above topics (and more) really need to be addressed sooner than later.


#27

Yes, you get realtime update with firestore, but it is a bit different. It is document based so you will get a notification when the whole document is updated. In Realm you can do a lot of transactions of very small objects and every change would be synced instantly. If you do structure your data in this way in firestore it would be very expensive as it is based on number of read/write operations, while Realm is not


#28

Ah. If you are concerned about that or want more granularity then the Realtime Database is for you. However, they’re rates are pretty generous; 50,000 reads per day for free and $.60 per Million reads after that.


#29

I don’t need the realtime updates, documents would probably be fine, but as I am already using Realm on both iOS and Android using Realm Mobile Platform would have been the ideal solution. I would just have to log in and then don’t have to worry about sync/backups.

I think this is getting a bit off topic, but I think we all agree that the current pricing is not very transparent and puts severe limits on what types of apps you can build using the platform


#30

Yeah - agreed. My driving point there though is the pricing model is pretty darn clear, and IMO fair. The pricing model scales with your user base and app.


#31

This thread is a bit scary and revealing to me. I went into a silo designing/MVPing my side project with full understanding that data/synch for the app will be Realm based starting with Realm Cloud. However it is quite scary that reading this thread could not find a link to a positive case study of Realm Cloud! Not a link to a Play Store/App Store of an app using Realm Cloud from an indie/small team. That is not encouraging.

Now that explains why Realm had to put themself up for acquisition. 350 paying customers out of 100,000 developers using Realm. That is not a great conversion rate and not looking promising. That is one more example against the Freemium business model. In a sense, an argument could be made that developers did not support Realm financially that is why some of these issues (Bugs, Incomplete Product Line, Lack of Support, Slow Response) ranted about in this thread exist.

I paid the $30 for a few months during development as my token of appreciation. I loved the product (initially before all these Platform noise), however with this uncertainty I will have no option than to go back to the old lady - SQLite via Room and cook up my own sync solution using Firebase Firestore.


#32

@Sipe A bit late answer, but yes, we need offline sync.

My (hopefully) last 2 cents to this thread. Continuing @valokafor thought:
Yes, 350 paying developers out of 100,000 is almost nothing.
I don’t mind to pay but a reasonable price.
I do not see much difference between $30 and $250 plan for me. The second one does not provide 8 times more services or resources. $250 plan still offers only 5x9 support.
Dedicated plans does not scale well and

The most reasonable pricing model would be ‘pay-as-you-go’ or pay for consumed resources.
Montly Active Device means that I should pay equally for the user who is actively consuming resources and the user who has just tried application once and uninstalled it.

By the way, I hope if we started to discuss Firebase it is not prohibited here to mention other products.
Does anybody give a try to Azure Mobile Apps? The pricing is complicated there but at least it should be reasonably scalable.

P.S. I have left a comment under MongoDB SEO post here https://www.mongodb.com/blog/post/mongodb-and-realm-make-it-easy-to-work-with-data-together though he didn’t noticed it :grinning:


#33

Answering myself: Azure Mobile Apps are mostly dead.


#34

I want to do a last comment to close this thread as well. I am on the Product side.

Your comments here have not gone unnoticed.
We agree that usage-based pricing does make more sense and are discussing how (and even more so when) to enable that.
This will be possible as we better integrate Realm with the capabilities that MongoDB provides (especially on the Atlas side). There is some work involved with that before this can be rolled out.

We believe that the resources available for the $30 account are pretty generous, thus you can run smaller scale applications on this tier. For purely development a lower tier would be beneficial. Something we are looking at as well (prior to rolling out usage based pricing).

We (and this includes MongoDB) are committed to make Sync a prime way for client/cloud communication. Much easier with the additional resources that MongoDB is putting into the project. Plus we now have additional resources to improve the Realm Client DB as well and address some of the outstanding issues faster. This should be good news for everyone.


#35

@robertherber
Thanks for your reply and explanation of the situation.
As far as I know from your Sales department you have stopped selling Dedicated plans at all until MongoDB decide on future licenses and prices.

  1. What are your recommendations for those who overgrow $30/month plan and need more reliability and resources to move to production?
  2. Are there any estimations regarding when the new pricing models will take effects or any decisions are taken? At least, approximately (1 month, 3 months, 6 months)?