Kenny Williams wrote a great article on this topic:

Since he has no comment capabilities on his website, I thought we can use this forum for discussions about it.

Personally, I would be interested to know what is the basis of this claim, because I have a hunch, that we might be experiencing a similar performance degradation:

As the number of databases grows, particularly if there are large databases, the performance of the system can degrade.
This is especially true if some databases dominate the system size, as it can impact the performance of all other databases.

I would also be interested to know what were the maximum number of Datomic Cloud databases in his experience, which were still sufficiently performant in his use-cases.

Hi @onetom,

Glad you found the article useful.

This statement comes from personal experience and discussions with the Datomic team :smiley: One of the systems I worked with using the by database tenancy approach hit on this exact case. The system was somewhere around 10b datoms with something like 60% of them residing in a single tenant. This caused major performance problems for all other tenants and use-cases served by this system. Fortunately, the performance problems were only limited to reads, so we were able to move that tenant and some other larger DBs into isolated query groups.

The article is intentionally not prescriptive with numbers because it is very dependent on your data and system behavior.

1 Like