Splunk Indexer Storage Sizing Practice

Free Splunk indexer storage sizing practice for the SPLK-2002 Architect exam: work out the cluster total, the average per peer node, the storage without a cluster or the minimum instances for an invented deployment, and Check my answer names the slip, such as multiplying everything by the replication factor.

Incoming data 
Retention 
Replication factor 
Search factor 
Indexers 

Check your answer to see the working.

Every question uses the Splunk ballpark: rawdata 15% and index files 35% of the incoming data. Splunk says real data can vary substantially.

Every deployment is invented. SmartStore, accelerations, summary indexes and space for the operating system are left out.

Practise the exams

Start with the free real exam questions for the SPLK-2002, then get the full pack. The SPLK-3003 pack is also useful.

SPLK-2002 Certified Architect
172 real exam questionsUS$39
Free: 43 questionsGet the full pack
SPLK-3003 Core Consultant
133 real exam questionsUS$39
Free: 33 questionsGet the full pack

Splunk storage sizing FAQ

The sizing rules in short
  • Ballpark: incoming data is stored at about 50%: rawdata 15%, index files 35%. Real data can vary substantially.
  • Data kept: GB per day x retention days.
  • Cluster total: rawdata x replication factor + index files x search factor.
  • Per peer node: the cluster total divided by the peers (the average).
  • No cluster: GB per day x days x 1/2, divided equally between the indexers.
  • Instances: at least replication factor + 2 (the peers, one manager node, one or more search heads).
How much storage does a Splunk indexer cluster need?

Take the rawdata (about 15% of the incoming data) times the replication factor, plus the index files (about 35%) times the search factor. The Splunk 10.6 example: 100GB of incoming syslog data with replication factor 3 and search factor 2 needs 15GB x 3 = 45GB of rawdata plus 35GB x 2 = 70GB of index files, 115GB across all peer nodes.

Why does replication factor 3 not need three times the storage?

Replicated copies of non-searchable data include only the data, not the index files. So with replication factor 3 and search factor 2, Splunk says you need more than two, but less than three, times the storage of non-clustered indexers.

Where do the 15% and 35% come from?

Both Splunk pages give them as a ballpark: after compression and indexing, incoming data takes about 50% of its original size, 15% for the rawdata file and 35% for the index (TSIDX) files. Splunk says this can vary substantially, so the drill states the ballpark in every question. You can measure your own data from a sample.

How much storage does each peer node need?

The Splunk examples divide the cluster total by the peers: 115GB on 3 peer nodes averages 38GB per peer, and 180GB on 5 averages 36GB. Splunk recommends the same disk storage on every peer, and each peer needs room for copies streamed from other peers too.

How do I size storage for indexers without a cluster?

Daily indexing rate x retention x 1/2. To keep 30 days at 100GB/day, plan at least (100*30/2) 1.5TB, and with 2 indexers divide it equally: 750GB each.

How many instances does an indexer cluster need?

At least replication factor + 2: replication factor number of peer nodes, plus one manager node and one or more search heads. With replication factor 3 that is five instances: three peers, one manager node and one search head.

What does the drill leave out?

SmartStore, data model accelerations, summary indexes, multisite clusters and the extra space for operating system disk checks and free space thresholds, which the Splunk example also leaves out. Every deployment is invented.

Is storage sizing on the SPLK-2002 exam?

The Splunk Enterprise Certified Architect test blueprint lists 3.2 Estimate non-smart store related storage requirements, 4.2 Identify disk storage requirements, 5.1 Identify non-smart store related storage and disk usage requirements and 14.1 Identify Splunk server roles in clusters.

Sources
  • Splunk Enterprise 10.6 documentation: System requirements and other deployment considerations for indexer clusters: the 15% rawdata and 35% index files ballpark, the two storage requirement examples (115GB and 180GB), non-searchable copies without index files, the same storage on every peer, and at least replication factor + 2 instances.
  • Splunk Enterprise 10.2 Capacity Planning Manual: Estimate your storage requirements: rawdata 15%, TSIDX 35%, the 30 days at 100GB/day example and dividing the space equally between indexers; How Splunk Enterprise calculates disk storage: daily indexing rate x retention x 1/2.
  • Splunk: Splunk Enterprise Certified Architect (SPLK-2002) test blueprint: 3.2, 4.2, 5.1 and 14.1.
  • Checked 21 November 2026. The tool's own approach: every deployment is invented; every question states the 15% and 35% ballpark, because Splunk says it can vary; cluster questions apply the cluster page sum to the data kept over the retention days; the per peer answer is the plain average, and any answer within 0.5 GB of it counts, as the page writes 38GB for 115GB on 3 peers. An independent re-solver that reads the ballpark from the saved pages and counts searchable and non-searchable copies agreed with the drill on 36,000 generated questions.

Checked 21 November 2026 against the Splunk Enterprise documentation for indexer clusters (10.6) and capacity planning (10.2) and the SPLK-2002 test blueprint.

More free tools

All free tools