[Question]: Graph storage performance - Postgres AGE is running slow

[acsangamnerkar]

Do you need to ask a question?

• [x] I have searched the existing question and discussions and this question is not already answered.
• [x] I believe this is a legitimate question, not just a bug or feature request.

Your Question

I have been playing around with the Graph storage and have used Azure Postgres - Flex server with AGE and VECTOR extension to address all the storage needs of Lightrag. I have also created the indices as documented in Lightrag documentation but the performance is extremely slow in the retrieval process. Even for top k of 10, with a graph of 3500 nodes and 4500 edges, it is taking 3-5 minutes for the response. The knowledge graph process is not an issue.

With support for Mongo dropping as a solution for all storages, there is a need to improve the postgres graph performance to increase the adoption of the lightrag as a solution.

Additional Context

Here is the postgres query as the longest running query by Azure monitoring:

SELECT query_sql_text FROM query_store.query_texts_view WHERE query_text_id=4153752584585438422; query_sql_text

SELECT * FROM cypher('chunk_entity_relation', $$ + MATCH (n:base) + OPTIONAL MATCH (n)-[r]->(target:base) + RETURN collect(distinct n) AS n, collect(distinct r) AS r+ LIMIT 1000 + $$) AS (n agtype, r agtype)

[danielaskdd]

PostgreSQL graph performance issue is on the priority waiting list.

[acsangamnerkar]

PostgreSQL graph performance issue is on the priority waiting list.

Thanks. I see the new branch created form the other issue.

[hamzafj5]

Hello, Is there any progress on this? As this is being eagerly waited on. Or if someone can help me in this, will be much appreciated.

[khizarhussain19]

Hi, Any updates on this ? I was able to run queries easily within 3-4 seconds with PGVector + Age when I loaded 2 documents only. However when the size of my documents increased to about 30, my CPU usage was becoming 100% and I was getting no response. The KG size is about 10k Nodes and 20k edges.

[danielaskdd]

PostgreSQL AGE performance issues have been resolved since v1.3.3. You can pull the latest version from the main branch and give it a try.

[danielaskdd]

However, for large-scale deployments, however, we recommend using Neo4j.

[danielaskdd]

PostgreSQL storage driver improved a lot, Please verify if the issue is resolved with the latest version.