Elasticsearch And Kibana
Summary: Elasticsearch is a backend search and analytics engine; Kibana is the web UI built on top of it. Together they're part of the Elastic Stack (ELK Stack).
At a glance​
| Question | Answer |
|---|---|
| What is it? | Elasticsearch = search/analytics engine; Kibana = its visual front-end |
| Why does it matter? | Clarifies where to query data (APIs) versus where to explore/visualize it (UI) |
| When do I use it? | Working with the Elastic Stack for search, logging, or analytics |
How it works​
- Elasticsearch stores, indexes, and searches large volumes of data quickly. It's backend-only — interaction happens via APIs/HTTP requests, no built-in UI.
- Kibana is the web application layered on top: browsing data (Discover), building charts and dashboards, and managing indices and settings.
- Both products are made by Elastic and ship together as the Elastic Stack.
Terminology vs traditional databases​
| Elastic term | Traditional DB equivalent | Description |
|---|---|---|
| Index | Table | Stores documents of a similar type |
| Document | Row / record | A single entry, stored as JSON |
| Field | Column | A key-value pair inside a document |
| Mapping | Schema | Defines each field's data type |
| Shard | Partition (roughly) | How Elasticsearch splits an index for scale/performance — see caveat below |
| Cluster | Database server | One or more nodes working together |
| Node | DB instance | A single Elasticsearch server in a cluster |
| Query | SQL query | Search/filter, written in JSON or ES|QL |
| Alias | View | A virtual name pointing to one or more indices |
Field types: text (full-text search), keyword (exact match), date,
integer/long, boolean.
Caveat: "shard = partition" is a useful first mental model, not an exact match. A shard is Elasticsearch's own unit of distribution and replication, which behaves differently from a strict relational-database partition — worth revisiting if you get into cluster sizing.
Remember​
- Elasticsearch is the engine/backend; Kibana is the UI on top of it.
- The term-mapping table is a starting intuition, not a 1:1 spec — shards especially.
Related notes​
- Kibana Data Views - how Kibana reads an Elasticsearch index
- Kibana Saved Searches - reusable filtered views built on a Data View