Learn how to optimize Elasticsearch and OpenSearch recovery with index recovery prioritization. Speed up critical index restoration and improve cluster availability after restarts.

When an Elasticsearch or OpenSearch cluster restarts—whether due to maintenance, a rolling upgrade, or unexpected downtime—it must recover its indices. But not all data is equal, and some indices are more important than others for getting your system back online quickly. That's where index recovery prioritization comes into play.

What Is Index Recovery Prioritization?

Recovery prioritization is a mechanism that lets you control the order in which indices are recovered after a cluster or nodes restart. Elasticsearch and OpenSearch assign a priority value to each index, determining which indices are recovered first.

The default priority is 1, but you can assign any integer value to influence the order:

"index.priority": 100

Higher numbers mean higher priority during recovery.

How to Set Index Priority

You can set recovery priority in the index settings:

PUT /important-index/_settings
{
  "settings": {
    "index.priority": 100
  }
}

Or during index creation:

PUT /critical-logs
{
  "settings": {
    "index.priority": 150
  }
}

The same thing applies to index templates of course.

When to Use It

Use recovery prioritization to:

  • Prioritize hot or frequently queried data
  • Accelerate recovery of system-critical indices (like Kibana, security indices, dashboards, etc.)
  • Defer recovery of archival or cold data to preserve I/O bandwidth

Elasticsearch (since v8.5) automatically assigns higher recovery priority to system indices like:

  • .security-*
  • .kibana_*
  • .fleet-*

These are critical for cluster functionality and user access, thus for many use-cases they should be recovered first.

Of course, you should avoid over-prioritizing too many indices—this defeats the purpose.

Don’t let critical data lag behind during recovery. With just a simple setting, you can make your cluster smarter about what gets restored first. Whether you're running a large-scale observability stack or a high-traffic search engine, index recovery prioritization is a small change with big impact.