In the Influx 1.8.7 version there are currently 28 databases. Want to know after ungrade to 2.7.5 whether we should create one bucket against each database or common for 2-3 databases.
If we create one bucket against each database, the count goes beyond 20 and on the InfluxDB portal it is mentioned as below:
A single InfluxDB 2.7 OSS instance supports approximately 20 buckets actively being written to or queried across all organizations depending on the use case. Any more than that can adversely affect performance.
We want to know what kind of performance impact are expected if the bucket count goes beyond 20.
Also if we create one bucket common for 3 to 4 databases what all problems we will face.
Each additional bucket in InfluxDB OSS 2.7.5 has separate caches and files on disk. As the number of buckets with active writes and queries increases, the opportunity for contention between the different buckets increases, particularly for memory and disk I/O. The 20 bucket recommendation is not a strict limit and the actual upper bound will be heavily influenced by the workload and hardware resources available. For example, it’s possible to run with hundreds of buckets with occasional writes/queries, but the risk of hitting a resource bottleneck is high.
That said, the way databases/buckets work is similar across InfluxDB OSS 1.8 and 2.7. If you are currently running InfluxDB 1.8.7 with 28 databases, you’ll probably be ok running the same number of buckets in InfluxDB 2.7.5.
