i don't want to speak for rest of Telemetry contributors but i don't think many(any) of us suggest using MongoDB or Ceilometer's API for storage. It is basically a data dump of what Ceilometer is collecting so it will be very, very verbose for most/all use cases.
as Joseph mentioned, Gnocchi[1] was developed to target billing/resource usage/health use case so it's optimised to store large sets of time-based measurements. Gnocchi v3 was just released so you may want to give that a try if you it fits your use case. i've been doing some performance testing on it this cycle that might help (specifically if using the Ceph backend)[2]. [1] http://gnocchi.xyz/ [2] http://www.slideshare.net/GordonChung/gnocchi-profiling-v2 On 14/09/2016 9:57 AM, Tobias Urdin wrote: > Hello, > > We are running Ceilometer with MongoDB as storage backend in production > and it's building up quite fast. > > I'm just having some simple thought on how large MongoDB setups people > are having with Ceilometer? > > More details about backup, replicas and sharding would also be appreciated. > > > I think we will have to look into moving our three replicas to sharding > in a short period of time. > > Best regards > > > _______________________________________________ > OpenStack-operators mailing list > [email protected] > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators > cheers, -- gord _______________________________________________ OpenStack-operators mailing list [email protected] http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
