Advanced Deployments - Implementing Splunk: Big Data Reporting and Development for Operational Intelligence - page 327

Advanced Deployments
When you first started Splunk, you probably installed it on one machine, imported
some logs, and got to work searching. It is wonderful that you can try the product
out so easily, but once you move into testing and production, things can get much
more complicated, and a bit of planning will save you from trouble later.
In this chapter, we will discuss getting data in, the different parts of a distributed
deployment, distributed configuration management, sizing your installation,
security concerns, and backup strategies.
Planning your installation
There are a few questions that you need to answer to determine how many Splunk
instances will be involved in your deployment:
• How much data will be indexed per day? How much data will be kept?
The rule of thumb is 100 gigabytes per day per Splunk indexer, assuming
you have fast disks. See the Sizing indexers section for more information.
• How many searches will be running simultaneously?
This number is probably smaller than you think. This is not the number of
users who may be using Splunk, but how many simultaneous queries are
running. This varies by the type of queries your group runs.
• What are the sources of data?
Where your data comes from can definitely affect your deployment.
Planning for all of the possible data that you might want to consume
can save you from trouble later. See the Common data sources section for
examples.