Ask HN: Largest DBs (any) you worked on?

ClickHouse cluster for web analytics data, 120 trillion records, mostly in two tables (hits and visits). 133 PiB uncompressed / 8.31 PiB compressed.

But it is not large compared to what other ClickHouse users have, for example:

> Ahrefs: main cluster is 100k+ CPU cores, 800TB RAM. 110PB nvme storage, uncompressed data size on main cluster is 1EB.

https://clickhouse.com/docs/en/about-us/adopters

Just a few hundred GBs of MySQL, which doesn't sound too big, but having tens of millions of rows can be tricky with some more complex query, to keep query time reasonable (within 1-2 seconds max).

~1.5 PB. Observability data.

size or disparate number of distinct dbs?

Size.

100s of TBs then. Nothing crazy.

Can you please more details on this database? How did you maintain and scale it? Issues with it something only type experience with it would learn? And things like that?

Oh yeah uhh not much to it. Just a lot of room on the company credit card and BigQuery. And lots of finding of tables that were just holding logs data and nuking those or at worst taking a subset and storing it in cold storage.

The rest is ad data of events and such. BigQuery is quite nice in that being managed we don’t have to do much.

It was 25

25Tb?

cubic meters