[meow_catrix]

The modern way is to sidestep the issue altogether and use Kubernetes with a database designed to run on Kubernetes. You can get sharding, replication and leader election essentially for free - you can concentrate on using the database instead of running the database.

Compute is really cheap compared to engineering man-hours.

[cpucycling7]

> The modern way is to sidestep the issue altogether and use Kubernetes

Kubernetes does require quite some time to learn/master. So you could say one replaces one time-consuming issue with another.

[kachapopopow]

https://artifacthub.io/packages/helm/bitnami/postgresql, postgres is fortunately one of them. Downside is that you cannot scale writes.

[remram]

What's "a database designed to run on Kubernetes"? Cassandra?

[linuxdude314]

Pretty sure they were referring to the operators you can install that will stand up production ready replicated clusters.

They work quite well.

[remram]

They replied to a comment mentioning "pgBouncer, repmgr, patroni" so supposedly not one of those.

[dgraph_advocate]

Dgraph is designed to run on Kubernetes.

[remram]

Not relational, not a replacement for Postgres.

Also doesn't look open-source at first glance.

[dgraph_advocate]

You should try it before claiming ”not relational”, as it can totally store and use relational data, and also be very useful.

Even though it says ”graph” you don’t need to write graph traversal queries - just describe in GraphQL the nested data model you want to get out of the database and that’s what you will get.

Community edition is open source.

[remram]

> Community edition is open source.

If that is true, you could definitely not make it less obvious. This is not the impression I got from your repo ("variously licensed under the Apache Public License 2.0 (APL) and the Dgraph Community License"). Neither your docs nor your website mensions a "community edition" either.

[dgraph_advocate]

It is not my project and I have nothing to do with Dgraph. After 15+ years of RDBMS, I just prefer it.

Also, first Google hit is dgraph.io:

> The only open source, AI-ready graph database that gives developers the tools to quickly build distributed applications at scale.

[fastest963]

Something like Yugabyte or Cockroach

[yrro]

Does the Cloud Native PG operator count?

[cheboygan]

+1 on checking out cloud native PostgreSQL operator, or other PG operators like crunchy or zalando or ongres

[dionian]

it was a minor pain finding and setting up a postgres operator in k8s, but once i got it going it wasn't too horrible. are these other solutions that are more built for it significantly easier to manage?

[slig]

Which one did you end up choosing?