Run an async ES|QL query

  • POST /_query/async
Auth

Asynchronously run an ES|QL (Elasticsearch query language) query, monitor its progress, and retrieve results when they become available.

The API accepts the same parameters and request body as the synchronous query API, along with additional async related properties.

Read the reference documentation

Required authorization

Index privileges read

Query parameters

allow_partial_results boolean · default: true

If true, partial results will be returned if there are shard failures, but the query can continue to execute on other clusters and shards. If false, the query will fail if there are any failures.

To override the default behavior, you can set the esql.query.allow_partial_results cluster setting to false.

delimiter string · default: ,

The character to use between values within a CSV row. It is valid only for the CSV format.

drop_null_columns boolean · default: false

Indicates whether columns that are entirely null will be removed from the columns and values portion of the results. If true, the response will include an extra section under the name all_columns which has the name of all the columns.

format enum

A short version of the Accept header, e.g. json, yaml.

csv, tsv, and txt formats will return results in a tabular format, excluding other metadata fields from the response.

For async requests, nothing will be returned if the async query doesn't finish within the timeout. The query ID and running status are available in the X-Elasticsearch-Async-Id and X-Elasticsearch-Async-Is-Running HTTP headers of the response, respectively.

Values: csv json tsv txt yaml cbor smile arrow

Request body

columnar boolean

By default, ES|QL returns results as rows. For example, FROM returns each individual document as one row. For the JSON, YAML, CBOR and smile formats, ES|QL can return the results in a columnar fashion where one row represents all the values of a certain column in the results.

filter {} Query

Specify a Query DSL query in the filter parameter to filter the set of documents that an ES|QL query runs on.

locale string
params union oneOf

To avoid any attempts of hacking or code injection, extract the values in a separate list of parameters. Use question mark placeholders (?) in the query string for each of the parameters.

One of: union[] or object[]
profile boolean

If provided and true the response will include an extra profile object with information on how the query was executed. This information is for human debugging and its format can change at any time but it can give some insight into the performance of each part of the query.

query string required

The ES|QL query API accepts an ES|QL query string in the query parameter, runs it, and returns the results.

tables map string to string to object

Tables to use with the LOOKUP operation. The top level key is the table name and the next level key is the column name.

include_ccs_metadata boolean · default: false

When set to true and performing a cross-cluster query, the response will include an extra _clusters object with information about the clusters that participated in the search along with info such as shards count.

wait_for_completion_timeout string · default: 1s

The period to wait for the request to finish. By default, the request waits for 1 second for the query results. If the query completes during this period, results are returned Otherwise, a query ID is returned that can later be used to retrieve the results.

keep_alive string · default: 5d

The period for which the query and its results are stored in the cluster. The default period is five days. When this period expires, the query and its results are deleted, even if the query is still ongoing. If the keep_on_completion parameter is false, Elasticsearch only stores async queries that do not complete within the period set by the wait_for_completion_timeout parameter, regardless of this value.

keep_on_completion boolean · default: false

Indicates whether the query and its results are stored in the cluster. If false, the query and its results are stored in the cluster only if the request does not complete during the period set by the wait_for_completion_timeout parameter.

Response

application/json
took number
is_partial boolean
all_columns object[]
name string required
type string required
columns object[] required
name string required
type string required
values union oneOf[] required
One of: number or string or boolean or object
_clusters object

Cross-cluster search information. Present if include_ccs_metadata was true in the request and a cross-cluster search was performed.

total number required
successful number required
running number required
skipped number required
partial number required
failed number required
details map string to object required
<string> object
status enum required
Values: running successful partial skipped failed
indices string required
took number
_shards object
total number required
successful number
skipped number
failed number
failures object[]
shard string required
index string required
node string
reason object required
type string required

The type of error

reason union string nullable

A human-readable explanation of the error, in English.

One of: string
stack_trace string

The server stack trace. Present only if the error_trace=true parameter was sent with the request.

caused_by object
Same fields as reason in failures
root_cause object[]
Same fields as reason in failures
suppressed object[]
Same fields as reason in failures
profile object

Profiling information. Present if profile was true in the request. The contents of this field are currently unstable.