Adding hosted clickhouse support + changing naming convention for cli…

[kakagri]

fixing issue with hosted clickhouse databases + naming issue, this is 100% vibe coded so be careful when merging :)

[vercel]

@kakagri is attempting to deploy a commit to the joshaavecom's projects Team on Vercel.

A member of the Team first needs to authorize it.

[jaggad]

To clarify this is because in hosted environemnt you want to just have default as the database and have tables be on the main db right

[jaggad]

Not sure we should bump this.

[jaggad]

Add these changes under the ## Changes Not Deployed section rather than making a new release.

This allows multiple pending changes to be deployed with a new version

[jaggad]

ClickHouse now uses single database from CLICKHOUSE_DB env var instead of requiring CREATE DATABASE privileges, rationale: giving all permissions to one script to drop any database is unreasonable.

Rationale makes a lot of sense to me, is a reasonable DBA perspective for permissions management. But this can be achieved by granting more granular permissions.

[jaggad]

Don't bump here

Suggested change

	version = "0.28.1"
	version = "0.28.0"

[jaggad]

Can we just call clickhouse.database_name instead of the getter fn

[jaggad]

Questions

@kakagri is the main purpose for this change specifically to create a user in a hosted environment that doens't need permissions granted to create and drop databases?

From what I understand there's no limitations in Clickhouse Cloud to create other databases (which act like Postgres schema's imo).

If you can help clarify the exact problem here I can maybe help, I've spent a while on reviewing this but more context on why this PR change is needed would be nice.

If tighter permissions are the goal here

CREATE DATABASE rindexer_internal;
CREATE DATABASE rindexer_internal_cant_drop;

CREATE TABLE rindexer_internal.test (id UUID) ENGINE MergeTree ORDER BY id;
CREATE TABLE rindexer_internal_cant_drop.test_cant_drop (id UUID) ENGINE MergeTree ORDER BY id;

CREATE USER jaggad IDENTIFIED WITH sha256_password BY 'password';

GRANT DROP DATABASE ON rindexer_internal.* TO jaggad;
GRANT DROP TABLE ON rindexer_internal.* TO jaggad;

Then try connected as the new user

DROP DATABASE rindexer_internal; -- drops ✅ 
DROP DATABASE rindexer_internal_cant_drop; -- doesn't drop, as expected ✅ 

Feedback

Test via examples to see remaining issues.

UniswapV3FactoryPoolCreatedPool::PoolCreated Event processing failed - id: bu5dd9OOvh - topic_id: 0x783cca1c0412dd0d695e784568c96da2e9c22ff989357a2e8b1d9b2b4e6b7118. Retrying... (attempt 8). Error: ClickhouseError: bad response: Code: 81. DB::Exception: Database rindexer_factory_contract_uniswap_v3_factory_pool_created_pool does not exist. (UNKNOWN_DATABASE) (version 25.7.3.13 (official build))

20 October - 11:51:26.601413 ERROR RocketPool::Transfer Event processing failed - id: OEzeDrgHOk - topic_id: 0xddf252ad1be2c89b69c2b068fc378daa952ba7f163c4a11628f55a4df523b3ef. Retrying... (attempt 3). Error: ClickhouseError: bad response: Code: 81. DB::Exception: Database clickhouse_indexer_rocket_pool does not exist. (UNKNOWN_DATABASE) (version 25.7.3.13 (official build))

@kakagri I think there may be a missed change somewhere in the creation for factory contracts.

Go to examples/clickhouse_factory_indexing and examples/rust_clickhouse and cargo run to see the errors printed in the console for testing.

---

Remaining unchanged locations

You'll still find lots of references to this in the codebase. E.g see the below insert_bulk used in some locations related to factory indexing.

.insert_bulk(
                &format!("rindexer_internal.{table_name}"),

Also checkout core/src/generator/events_bindings.rs and other files in the codegen section which are fired when rindexer codegen indexer is used and codegens like the following but you've changed it from this layout clickhouse_indexer_rocket_pool.transfer to this layout clickhouse_indexer_rocket_pool_transfer.

let result = context
                .database
                .insert_bulk("clickhouse_indexer_rocket_pool.transfer", &rows, &postgres_bulk_data)
                .await;