PlanetScale Scaler Pro
planetscale.comI love PlanetScale and if I had one request it would be a plan between Hobby and Scalar (or even a way to combine multiple DBs in 1 plan, and no 2 prod branches won't cut it).
I have 2 paid Scalar DBs on PlanetScale and I have no intention of moving elsewhere but it does kill me that they both sit almost unused ~10 months out of the year (I have bursty traffic and only during the events, in-person events, that the software is built for). At ~$348/yr per DB it's still a steal compared to managing it all myself but I look at my usage (even during my "busy" months) and I barely make a dent in the usage tier I'm on. In fact I think you could total up my total usage for the lifetime of my account (both DBs) and they wouldn't total up to 1 month of the usage tier.
Again, I'm not complaining and the cost is manageable but I did create and sell some new software in the last year that I built on DynamoDB (in part to learn, in part due to costs). My software that uses PS is single-tenant so I need 1 per client which is on me, if I was able to rewrite it to be multi-tenant then I'd only have to pay $348 total a year instead of per-client.
All in all I have had nothing but good experiences with PlanetScale from the product itself to the support staff. I love the migrations and the rollback support, it feels natural when you start using it and dealing with migrations in other DBs feels like a huge pain once you've done it in PlanetScale.
I'm interested in using PlanetScale.
Would you happen to know the magnitude of how many simultaneous connections could $348/yr buy me in PlanetScale?
One of my clients hovers around 50 simultaneous connections to their main MySQL server on a normal day, but they have bursts of 3k simultaneous connections for an entire day, twice per year.
Their workload is about 5% writting (INSERTs, UPDATEs), 95% reading (SELECTS).
Planetscale advertises up to 10,000 concurrent connections on their $29/mo Scalar plan (https://planetscale.com/pricing - click on "Compare all plan options and benefits"). Being based on Vitess as well, I believe that number.
The blog post shows both the Scaler and Scaler Pro plan supporting up to 10,000 concurrent connections: https://planetscale.com/blog/announcing-scaler-pro#scaler-pr...
This number of connections on any of the paid plans should be no issue.
you might find Supabase[0] is closer to what you need. I'd recommend giving them a look, that is, if MySQL compat isn't a requirement of course, since Supabase is all PostgresSQL.
disclosure: not a employee or investor in Supabase, but I sure am a fan.
I kind of have a thing for DBaaS pricing, and that table with the comparison with RDS looked suspicious to me because it doesn't specify the exact instance type used on the AWS side. ~~I think it should be `db.r6g.xlarge` because it has 4 vCPUs and 32 GB RAM. That is $0.43/h, so $0.43 * 730 = $313 / month. They have one primary and 2 replicas, so 313 x 3 = $940. It doesn't quite fit, so maybe that's not the instance type they chose, but indeed it seems cheaper.~~ But that doesn't take into account reserved instances, which can lower the price on the RDS side.
Edit: Looking again, I think the instance used for comparison is `db.r6gd.xlarge` from the Multi AZ-deployments (two standbys) list. That is $1.445/h, so $1054 / month. The difference could be for storage and I/O.
However, the PS Scaler Pro is $1.5 / GB, which is quite a lot. General purpose storage in AWS is only $0.115. The comparison table uses 10 GB only, but if the DB size is 1 TB, then RDS would be a lot cheaper?
Please correct me if I got something wrong, I'm sure there's stuff I'm missing.
If they are running on AWS (which it looks like they are), they aren’t going to be able to undercut AWS. For a retail user who doesn’t want to commit to RIs or have enough volume for an EDP, sure PS can be cheaper.
The value I found is being able to do multi-region read replicas with no compute overhead for lower traffic geos.
I like the idea of PS and have toyed around with the idea of migrating to it but there are some glaring issues I don’t want to deal with:
- no native way to export backups and avoid vendor lock in (or pay for the row reads to generate regular backups)
- contradictory cost model. Their pricing page reads “Every time a query retrieves a row from the database, it is counted as a row read. Every time a row is written to the database, it is counted as a row written.” while their docs state “Rows read is a measure of the work that the database engine does, not a measure of the number of rows returned. Every row read from a table during execution adds to the rows read count, regardless of how many rows are returned.”
That’s not contradictory?
It’s very common for scale-out architectures to read more data than is ultimately returned, because the former is pulled from individual shards and then some centralized filtering / post processing is applied in some API middleware layer.
Trying to fix that by pushing down more of the query/execution is sometimes but not always feasible or practical.
(responding to GP) “Full table scans are expensive” is a better way to reframe the pricing model (at least before this pricing change). The distinction is between rows _examined_ vs rows _returned to a db client_. Even the latter is a tricky concept with vitess since it’s a routing/sharding system that sits on top of vanilla MySQL instances.
> PS Scaler Pro is $1.5 / GB, which is quite a lot. General purpose storage in AWS is only $0.115.
GP storage is pretty slow (relatively). I hope PS isn't giving us that out of the box and does provision more IOPs etc.
AWS bills for absolutely everything as PS mentioned in the article. This can include temporary files and what not that can take up a large chunk of your storage.
> because it doesn't specify the exact instance type used on the AWS side
That in a way is problematic because CPU generations matter and even if PS is using the "best" currently, how do we know if they'd keep up? Prices could swing the other way quite easily.
Am I missing something in the pricing or is the Storage costs very unappealing. For anything with a usage skewed towards high ratio of storage to usage, the costs aren't competitive.
2.5$ per GB in "Scalar" and 1.5$ per GB in "Scalar PRO" compared to 0.11$ RDS for General purpose (or 0.125$ for provisioned + the IOPS you use, double that for multi-az), Supabase at 0.125$, Firebase at 0.1725$,DynamoDB at 0.25$, MongoDB Atlas serverless at 0.25$, cockroachDB serverless at 0.5$ per GB, FaunaDB at 1$ per GB. (Neon says it's 0.000164 per GiB, but somethings seems off, it's not at the same scale, so I'm guessing there's a catch here)
> Am I missing something in the pricing or is the Storage costs very unappealing
Yes, there's no comparison on the IOPs or anything else and what counts as storage. They all count slightly differently and so unless you account for it all it's a bit moot.
AWS alone has different tiers of storage from HDD to NVME ssd. The pricing varies greatly.
RDS will literally count anything including temporary files as usage of storage. Planetscale claims not to.
Neon price is listed as per hour, instead of per month like the others :)
While you're not necessarily missing something, it's worth pointing out that usually storage costs don't dominate the bill, the compute costs are usually higher. But with the PlanetScale storage price being more then 10x the price of the other, it's definitely something to keep in mind, and it can dominate the bill.
Excited to see this finally rolled out. Now just waiting on the foreign key constraint support.
EDIT: Ignore this comment, I was mistaken
I don't believe that will ever happen due to the underlying Vitess tech [0] that PlanetScale uses
We are working on foreign key constraint support right now. :)
It's going to happen.
Very cool! I didn't know that was on the roadmap.
It's on the official Vitess roadmap, though you do have to search on this page for it https://vitess.io/docs/resources/roadmap/
It's on its way!
This is great news. I strolled around https://github.com/vitessio/vitess/issues/12967.
Are there any public discussions of more trade-offs vitess has to make to enable fks?
This is the only thing that is stopping me from trying PlanetScale. I can't imagine not having the database do this work for me
This is the only real thing stopping me from trying PlanetScale as well, glad to see in the comments that it's coming.
It's great MySQL is still getting some love. I have no complaints with it.
What is the expected performance? If I have a very simple table with 100k records, what will be the expected read/write/update tps? I know it depends on many things, but still any guesses?
Tl;dr: This new plan basically fix the number one complain about PlanetScale and offers Unlimited row reads and writes.
Next up is the new number one complain, lack of PITR. When that's ready I can finally realistically propose moving to PlanetScale to my team. I cannot wait for the day.
I'd love to learn about your requirements. Can we chat? s@planetscale.com
wow that's incredible!!
The only thing stopping me from using this is the lack of foreign keys. Is that ever coming?
Yes, it's on its way.
Would be nice for the first table to have the pricing, and a comparison with the old plan as well...
The pricing page shows the old plan starts at 29$/month for row-based pricing, and the new one at 39$/month for resource-based pricing: https://planetscale.com/pricing
We are getting the chart updated with the pricing now. Thanks for pointing that out.
love PlanetScale product. Do you recommend any similar offering for PostgresSQL?
Neon is a great suggestion. I'm just writing to add a bit more color.
Neon is a bit different from PlanetScale. With PlanetScale, they're running Vitess under the covers. Vitess is a proxy that sits between you and the MySQL database. It means they can do fancy things like rewrite the queries on the way to the database or route them differently. For example, a "SELECT" query could be sent to a replica instate of the primary. Likewise, if you've sharded/partitioned your data by customer_id, Vitess could see the query says "WHERE customer_id = 5" and send it to the correct server. This proxy also means that Vitess can seamlessly manage some things for you like failing over to a new database or bringing online a new replica.
Neon is a decoupling of storage from the database processing. While Vitess still has MySQL storing the data, Neon changes the storage so that it can separate data processing and data storage. This is somewhat like Amazon's Aurora.
If you're interested in other PostgreSQL compatible options, CockroachDB is PostgreSQL compatible, but is more of a fully distributed database from the ground up rather than new layers like Neon or PlanetScale.
I work at Neon like the other person mentioned, but I believe Supabase is Postgres compatible, though I haven't looked into them much.
The last time I recommended to use planetscale at my company the only factor that stopped us using them was GDPR/DSGVO. Any news about a europe friendly version?
Do you know specifics they're missing?
At a glance, their Data Processing Addendum[0] seems to address GDPR, along with country-specific regulations like Swiss DPA and UK GDPR