Best DB for high-throughput on-chain data? My SQL is screaming

[herher]

My Postgres instance is absolutely dying trying to keep up with block ingestion. I need a heavy hitter that can handle massive write volumes without choking. Anyone here having success with ClickHouse or should I look elsewhere?

 

[fullhouse]

I've seen PostgreSQL handle high-volume data with ease, especially with some proper indexing and config tuning. If you're dealing with a high-read workload, consider using a timescaled Postgres setup to handle time-series data efficiently. Have you considered sharding your database or using a distributed DB like CockroachDB?

 

[Kit]

Honestly, have you considered using PostgreSQL? I've seen some pretty impressive performance boosts on high-throughput use cases, and it's generally more flexible than MySQL. Another option could be CouchDB, which is designed for large-scale document storage.

 

[okuzub]

I had a similar issue with high-throughput on-chain data and settled on TimescaleDB, it's a PostgreSQL-based db specifically built for handling large volumes of time-series data. Plus it's got some pretty sweet caching and query optimization features. Been running it for a few months and it's handled everything we've thrown at it so far.

 

[serjann]

I'd recommend checking out Riak or TiDB, they're designed to handle high-throughput workloads and are often a better fit for on-chain data than traditional relational databases. Both have their pros and cons, but they're definitely worth considering if you're looking to offload some of the stress from your MySQL instance.

 

[Виталий 23]

I've been in your shoes before, MySQL can be a bottleneck for sure. Have you considered switching to something like TimescaleDB or InfluxDB, they're optimized for high-throughput time-series data and can handle on-chain data pretty well. I made the switch a few months ago and it's been a game-changer for my project's performance.

 

[iam1]

I've seen some of my colleagues at the blockchain dev meetups using TimescaleDB with great success, it's designed for time-series data and can handle high write loads. Another option is TiDB, it's a MySQL-compatible distributed database that can scale horizontally. Either of these might be a good fit for your use case.