SURVEY: Who is using Dgraph?

Moved from GitHub dgraph/5567

Posted by vvbalaji-dgraph:

Hey there :wave:, we’re doing an informal survey to gauge the interest in this project and it would be very helpful if you could answer the following questions:

  • If applicable, what company/organization do you represent?
  • What other databases did you evaluate/consider before picking Dgraph?
  • What kind of problems are you solving with Dgraph?
  • Can you provide any information about your cluster configuration and amount of data that you store in Dgraph?
  • If you are not using Dgraph but chose another database, what were the reasons?
  • If applicable, please share a link to your project/product built with Dgraph.

Thank you for your help in improving Dgraph!

TiemenSch commented :

How would you want us to answer this? I’m still more or less evaluating Dgraph and trying to get it to work with our own internal software/services, but leaving the answers to all of the above questions on a public issue does not seem ideal.

nicklanng commented :

I’m evaluating dgraph for a startup right now. Currently using Neo4j but we have problems with scaling and intermittent issues.

Dgraph looks great. I’ve got a proof of concept up and running really quickly and loving the performance. The primary use case will be for calculating best paths between nodes and ABAC style access controls.

However, having features like ACL and backups behind such a high paywall are definitely a problem and may have us looking elsewhere.

grizzly-monkey commented :

We wanted to evaluated dgraph for our one of the microservices where we needed heavy lifting of graphs and tree structured data in Industrial materials management.

We have around 80K nodes on average in one tree and with this neo4j starts complaining on memory. (as you know its Java)

Dgraph was performant , intuitive , and reliable , however as the open soruce did not have multidb support so we went with using Neo4j.
If Dgraph brings this feature we are surely moving to dgraph. for all our materials management workloads.

Willem520 commented :

I have used Dgraph in my business. it made us amazing.but there is some problems are bothering me.

  1. dgraph could not support multi database or tenancy in one cluster, it means one business must need one cluster, that makes hardware cost is too high.
  2. I have a large dataset about 50 billion. it is difficult for us to load it into cluster by bulk loader or live loader(because of oom).
  3. I want to mutate data in realtime by pipeline like kafka or other streaming processing.currently, dgraph cant not support high throughput very well.
  4. I do some tests about insert data and modify data. it got lower query throughput after inserting data.I had a discussion(Query throughput),but from now,it is unresolved.
  5. Now, Dgraph can not support bigdata system.So, it means if I have a large dataset in hive, hdfs and so on, there is no way to load dataset into Dgraph conveniently.

because of the above major obstacles, it is hard for us to large-scale use dgraph in production.

amaster507 commented :

  • If applicable, what company/organization do you represent? IFBMT/MissionBase
  • What other databases did you evaluate/consider before picking Dgraph? MySQL, MariaDB, DynamoDB, MongoDB, RDS, Neptune, Neo4j
  • What kind of problems are you solving with Dgraph? the n+1 problem and removing the API layer from our stack.
  • Can you provide any information about your cluster configuration and amount of data that you store in Dgraph? When in Triple format ~420Mb which is ~6.8M triples.
  • If you are not using Dgraph but chose another database, what were the reasons? N/A some team members want to use Netpune but we would love to stay on open source projects
  • If applicable, please share a link to your project/product built with Dgraph. It will hopefully soon be on https://missionbase.com