Question. Graph-node deployment takes to much space

[statzky]

Hello everyone, I'm sorry if this is not the right place to ask..

I have a graph-node deployed on a virtual machine (in docker), and I'm encountering an issue with disk space consumption. My setup details are as follows:

• Network: Polygon
• Start block: 2,000,000
• Current block: 59,606,297
• Database size after initial indexing: 35 GB (on 58,908,274 block)

The problem is that the database size increases by approximately 7 GB every day. Given the number of new blocks generated each day, this seems unusually high.

Here are some additional details:

• I have 5 subgraphs deployed to my node.
• Subgraphs are tracking events from certain smart contracts.
• Using PostgreSQL for the database (nothing special, just like all other subgraphs).
• Lately I have no new transaction for my contracts (around few weeks).

I suspect that the rapid increase in disk space usage might not be solely due to new blocks and their associated data. Some potential culprits or considerations might be:

• Node configuration (everything is set by default, maybe I should change something?).
• Database indexing or maintenance issues.
• Possible redundant data or "bloat" in the database.

I know about GRAPH_ETHEREUM_CLEANUP_BLOCKS, but I want to use my graph-node for production purposes.

Should I use Postgres VACUUM or ANALYZE functionality?

Could anyone provide insights or suggestions on what might be causing this rapid increase in database size and how to optimize it?