Memory requirement for graphql node setup

Hello,

How much maximum memory will require for graphql node setup if followed the same steps given here input-output-hk/cardano-graphql with docker setup?

1 Like

I’m a bit late here, but I’ll leave what I’ve found out so it might help others.
I’ve been working with the GraphQL set of docker services, and I’m now sure, you cannot finish db sync with less than 16 Gb Ram.

I Don’t believe there’s a MAX value here, but I can assure you it wont finish sync with 8GB Ram, the max seems to depend on the size of the chain, which seems unsustainable (at current state 4.0.0), as from my observations, it seems that the bigger the chain, the more it needs to hold in memory to complete the sync.

It’s impressive how demanding this set of services (hasura/db-sync/graphql/postgres/node) is during sync, it definitely cannot run on low spec machines which is not sustainable long term, but I’m confident a few versions down the road, it will improve, there are some ongoing discussions over github issues around resource usage.

It would have saved me HOURS, if we had a recommendation on specs on the Github Page.

Hello,

Thanks for your reply.

I’ve set up a node on VM with 64GB RAM and 8 vCPU. Node was syncing fine previously but from 2 days cardano_graphql service is restarting intermittently. I’m not sure about the reason.
I’ve seen cardano_db_sync_extended service hitting 100% sometimes. Any idea is that the reason for it?

Here are the screenshots,

Can anyone please help here?