ObservKit / Capacity Planning / Platform comparison
Capacity planning

Compare platform storage

One daily ingest, four platform models: Splunk, Elasticsearch, Grafana Loki and OpenSearch side by side, each with the defaults of its own calculator.

● Calculated locally in your browser
Your scenario
Start with what you know
Monthly storage cost optional, your prices

ObservKit has no built-in prices. Enter your own (any currency); the estimate covers storage capacity only, not compute, requests or licences.

Same input, four models

Not a ranking Each card is a different quantity from its own model. Open a card to refine it in that calculator with the same input.

Grafana Loki

Replication factor 3 · 6 ingesters

Object storage dataset, one copy

28.1 TB

Ingester WAL, cluster-wide: 17.4 GB–26 GB

Open in the Loki calculator

Platform-neutral baseline (raw volume × 2 copies + growth, no compression): 56.3 TB · Open in Log Storage

How to read the comparison

  • Splunk stores rawdata on every copy (× replication factor) and index files on searchable copies (× search factor), on the indexers’ local disk.
  • Elasticsearch splits retention across hot, warm and cold tiers, adds one replica and divides by an 85% planning utilization of local disk.
  • Grafana Loki keeps one copy of compressed chunks in object storage; replication affects the ingesters and their WAL, not the object store.
  • OpenSearch follows the Amazon OpenSearch Service sizing guidance: one replica, an index factor of 1.1 and the service’s storage overhead.

These are planning estimates for storage capacity. They do not include compute, licences or query load. Why the numbers differ →

Refine one platform

The comparison uses each calculator’s defaults so the four models stay comparable. Replication, tiers, shards, SmartStore and compression ratios are set in each calculator, where every assumption has its source.

Methodology · Guides