Elasticsearch Elasticsearch Long Running Shard Tasks

By Opster Team

Updated: Mar 10, 2024

| 2 min read

What does this mean?

A long-running shard task in Elasticsearch refers to a task that is taking an unusually long time to complete. Shard tasks are operations performed on individual shards, such as indexing, searching, or relocating. When a shard task takes longer than expected, it can be considered a long-running task.

Why does this occur?

Long running shard tasks can occur due to various reasons, such as:

  1. High query load: A large number of queries or complex queries can cause shard tasks to take longer to complete.
  2. Insufficient resources: If the Elasticsearch cluster does not have enough resources (CPU, memory, or disk space), shard tasks may take longer to complete.
  3. Slow or unresponsive nodes: If a node in the cluster is slow or unresponsive, it can cause shard tasks to take longer to complete.
  4. Large shard size: If a shard is too large, it can take longer to perform operations on it.

Possible impact and consequences of long running shard tasks

The impact of long running shard tasks can include:

  1. Reduced cluster performance: Long running shard tasks can consume resources and slow down other tasks in the cluster.
  2. Increased latency: As shard tasks take longer to complete, the overall response time for queries may increase.
  3. Potential data loss: If a long running shard task is related to replication or recovery, it may result in data loss if not resolved in a timely manner.

Resolving long-running shard tasks

How to resolve long-running shard tasks in Elasticsearch?

  1. Monitor and optimize queries

    Analyze the queries causing long-running shard tasks and optimize them to reduce their execution time.

  2. Allocate sufficient resources

    Ensure that the Elasticsearch cluster has enough resources (CPU, memory, and disk space) to handle the workload.

  3. Fix non-cancellable and long-running shard tasks

    Identify and fix shard tasks that are non-cancellable and long-running to prevent them from affecting cluster performance.

  4. Balance shards during off-peak hours

    Reroute shards, add or remove new nodes, and perform other shard balancing operations during off-peak hours to minimize the impact on cluster performance.

To identify long running tasks, use the following command:

GET /_tasks?detailed=true&timeout=30s

To cancel a specific task, use the following command:

POST /_tasks/<task_id>/_cancel

To reroute a shard, use the following command:

POST /_cluster/reroute
  "commands": [
      "move": {
        "index": "<index_name>",
        "shard": <shard_number>,
        "from_node": "<source_node>",
        "to_node": "<destination_node>"


Long running shard tasks can affect the performance of an Elasticsearch cluster. By understanding the causes and potential impacts of this event, you can take appropriate steps to resolve the issue and maintain optimal cluster performance.