When using VictoriaMetrics, the required storage space depends on the
number of plugins enabled in the postgrespro receiver
and the number of monitored DBMS objects. The more plugins enabled and the
more objects monitored, the more metrics will be generated, and the more
storage space will be required for VictoriaMetrics.
The required disk space for VictoriaMetrics in typical scenarios is provided below. All scenarios use the cluster configuration with three instances and a replication factor of 2.
Collecting metrics from a basic set of plugins requires 6 MiB of daily storage per DBMS instance when stored in the VictoriaMetrics cluster.
Plugins enabled:
hostmetrics: cpu,
disk, filesystem,
load, memory,
network, paging,
processes
postgrespro: activity,
archiver, bgwriter,
cache, databases,
io, locks,
replication, replication_slots,
tablespaces, version,
wal
The data size for the plugins listed above does not depend on the number of objects within the DBMS instance.
Collecting metrics from an extended set of plugins will require 143 MiB of daily storage per DBMS instance when stored in the VictoriaMetrics cluster.
Plugins enabled:
hostmetrics: cpu,
disk, filesystem,
load, memory,
network, paging,
processes
postgrespro: activity,
archiver, bgwriter,
bloat_indexes, bloat_tables,
cache, databases,
indexes, io,
functions, locks,
replication, replication_slots,
tables, tablespaces,
version, wal
Number of objects (tables, indexes) in the DBMS instance: 10,000.
The data size for the bloat_indexes, bloat_tables, databases, indexes, functions, and tables plugins depends on the number of objects within the DBMS instance.