8 Replies
Yes. Autoscaling is an option of creating a cluster. Beyond that, clusters provide better isolation and performance. You can read the gist here. https://xata.io/docs/dedicated-cluster
Dedicated clusters
How to manage your own Postgres instance
si it won’t autoscale if i don’t pay at least 50$ a month (for a cluster) ?
By default, the shared clusters will scale fine. If you have more serious needs, and require isolation, then dedicated clusters provide a more cost effective option for high traffic requirements. If your usage is pretty steady, an non-autoscaling cluster is best, because you can just pick what you need. If you require true autoscaling, where the cluster itself changes based on load (volatile, unknown traffic) that an autoscaling cluster is best. Xata allows you to move branches between each option once you have them set up, so it's pretty easy to move things around if you need.
and did you have any info on max read/write transaction per sec it can handle ?
hi, on the free and the pro shared clusters we enforce concurrency limits (which apply over the wire protocol and data api) and rate limits (which only apply for the data api, and sql over http). You can see the values and definitions here:
* Concurrency limit: https://xata.io/docs/concepts/pricing#concurrency-limit
* Rate limit: https://xata.io/docs/concepts/pricing#rate-limit
Let me know if you have any questions. In our experience, the above are large enough to not impact any light application. If you hit those limits, it's likely fair and expected to pay for a dedicated cluster.
Pricing
Xata provides transparent and flexible billing options to accommodate user needs
ok so if i have more than 15 concurrent users i can't use your product
well it look like i will have to build my own database provider then
you'd typically want to use a connection pool in your application. What language / framework do you use?
i don't this how this will help my app with a consistent 10req/sec