← All insightsProduction

Running Elasticsearch in production: a hardening checklist

CodeWizard Engineering · 13 August 2025 · 3 min read · Advanced

Demos and production are different sports. This checklist covers what changes when Elasticsearch has to be reliable, secure and observable under real load.

Security and reliability defaults

  • Enable security (TLS + auth); never expose 9200 publicly
  • Use role-based access and API keys

A production-minded config

Create an index with an explicit mapping.

bash
curl -X PUT localhost:9200/logs -H 'Content-Type: application/json' -d '{
  "mappings": { "properties": {
    "@timestamp": { "type": "date" },
    "level": { "type": "keyword" },
    "message": { "type": "text" }
  } }
}'

Operational best practices

  • Define explicit mappings; disable dynamic where possible
  • Use Index Lifecycle Management for retention/rollover
  • Size shards deliberately (tens of GB each)
In production, boring and reproducible beats clever every time.

Want senior help putting Elasticsearch to work - or a team bootcamp on it? CodeWizard runs Israel's largest DevOps community and embeds named senior engineers with a transparent hours log. Book a free consult or explore the Academy.

Like where this is going?

Get hands-on help from the engineers who write these posts.

Book a free consult