Too many things come into play here, the only way to know is to test both approaches on your use case. By the way, do you have the same table set for every tenant? It may be beneficial to group caches not by tenant, but by data size and type.
On Thu, Sep 26, 2019 at 12:56 PM siva <[email protected]> wrote: > Hi Pavel, > > I have some queries related to cache groups ,if you help it will be useful > for us. > > > *instead of one cache per company, use one cache group per company. > * > 1.What is the performance impact > =>Is there any number of cache group limitations? > here cachegroups proportional to tenants > > 2.What If we have only cluster of 3 nodes(each node 16cpu and 56GB RAM with > ssd) handling 20 cachegroups with each cache group contains 120 caches(i.e > 2.4k caches aprox) > > > *That said, they might impact the performance of read operations and > indexes > lookups > *is it reasonable or drastically decrease? Since Our application is for > both > read and create/update intensive (before create or update we should > query/read data from different tables) > > > > > Thanks > siva > > > > -- > Sent from: http://apache-ignite-users.70518.x6.nabble.com/ >
