graphite-dev team mailing list archive

[Scott Smith <question177524@xxxxxxxxxxxxxxxxxxxxx>]

Question #177524 on Graphite changed:
https://answers.launchpad.net/graphite/+question/177524

Scott Smith posted a new comment:
FYI, I've got a cluster of three hosts with two 200GB SSDs each (JBOD),
running four carbon-caches each with consistent hashing and a
replication factor of 2, so each data point is stored in two instances.

I just checked the carbon stats and metricsReceived is peaking at around
700k each day. I expect this to break 1M in the coming weeks.

I'm not storing everything I could yet, as I had to scale graphite twice
already. First I was I/O bound, and I solved that with an SSD. Then once
I started rolling out host metric collection of literally everything I
could record (redis, postgres, cpu, df, iostat, memory, network,
tcp/udp, etc etc.) I quickly ran out of storage capacity.

So now I've had this cluster up for a couple weeks and things are going
quite well. Each drive is at around 30% used and a total of about 2400
writes/sec across the cluster.

-- 
You received this question notification because you are a member of
graphite-dev, which is an answer contact for Graphite.