Link Search Menu Expand Document Documentation Menu

CAT APIs

Introduced 1.0

You can get essential statistics about your cluster in an easy-to-understand, tabular format using the compact and aligned text (CAT) API. The CAT API is a human-readable interface that returns plain text instead of traditional JSON.

Using the CAT API, you can answer questions like which node is the elected cluster manager, what state the cluster is in, how many documents are in each index, and so on.

Example

To see the available operations in the CAT API, use the following command:

GET /_cat
response = client.cat.help()

The response is an ASCII cat (=^.^=) and a list of operations:

=^.^=
/_cat/allocation
/_cat/segment_replication
/_cat/segment_replication/{index}
/_cat/shards
/_cat/shards/{index}
/_cat/cluster_manager
/_cat/nodes
/_cat/tasks
/_cat/indices
/_cat/indices/{index}
/_cat/segments
/_cat/segments/{index}
/_cat/count
/_cat/count/{index}
/_cat/recovery
/_cat/recovery/{index}
/_cat/health
/_cat/pending_tasks
/_cat/aliases
/_cat/aliases/{alias}
/_cat/thread_pool
/_cat/thread_pool/{thread_pools}
/_cat/plugins
/_cat/fielddata
/_cat/fielddata/{fields}
/_cat/nodeattrs
/_cat/repositories
/_cat/snapshots/{repository}
/_cat/templates
/_cat/pit_segments
/_cat/pit_segments/{pit_id}

Optional query parameters

The root _cat API does not take any parameters, but individual APIs, such as /_cat/nodes, accept the following query parameters.

Parameter Description
v Provides verbose output by adding headers to the columns.
help Lists the default and other available headers for a given operation.
h Limits the output to specific headers. Accepts column names, column aliases, and wildcards (for example, routing.*).
format The format in which to return the result. Valid values are text (default), json, yaml, cbor, and smile.
s Sorts the output by the specified columns. Accepts column names and column aliases, each optionally followed by :asc or :desc (for example, s=a:desc). Each sort key must be an existing column name or alias.
limit The maximum number of rows to return. Must be greater than 0. Applied after sorting (s). For more information, see Limit the number of rows.
bytes The units used to display byte values. Valid values are b, kb, mb, gb, tb, and pb. For more information, see Supported units.
time The units used to display time values. Valid values are nanos, micros, ms, s, m, h, and d. For more information, see Supported units.
ts Boolean. Whether to include timestamp columns in operations that provide them, such as _cat/health. Default is true.
pri Boolean. Whether to add primary shard statistics in operations that provide them, such as _cat/indices. For each column specified in the h parameter that has a primary shard counterpart, adds a pri. column. For example, h=i,ss with pri returns the i, ss, and pri.ss columns.

Specify these query parameters in any CAT operation to obtain more specific results. The following examples use the CAT aliases operation.

Get verbose output

To query aliases and get verbose output that includes all column headings in the response, use the v query parameter.

GET /_cat/aliases?v
response = client.cat.aliases(
  params = { "v": "true" }
)

The response provides more details, such as names of each column in the response.

alias         index          filter routing.index routing.search is_write_index
.kibana       .kibana_1      -      -             -              -
sample-alias1 sample-index-1 -      -             -              -
sample-alias2 sample-index-2 -      -             -              -

Without the verbose parameter, v, the response contains the same columns without the header row:

.kibana       .kibana_1      - - - -
sample-alias1 sample-index-1 - - - -
sample-alias2 sample-index-2 - - - -

Get all available headers

To see all the available headers, use the help parameter:

GET _cat/{operation_name}?help

For example, to see the available headers for the CAT aliases operation, send the following request:

GET /_cat/aliases?help
response = client.cat.aliases(
  params = { "help": "true" }
)

The response contains the available headers:

alias          | a                | alias name
index          | i,idx            | index alias points to
filter         | f,fi             | filter
routing.index  | ri,routingIndex  | index routing
routing.search | rs,routingSearch | search routing
is_write_index | w,isWriteIndex   | write index

Get a subset of headers

To limit the output to a subset of headers, use the h parameter:

GET _cat/{operation_name}?h={header_name_1},{header_name_2}

For example, to limit aliases to only the alias name and index, send the following request:

GET /_cat/aliases?h=alias,index
response = client.cat.aliases(
  params = { "h": "alias,index" }
)

The response contains the requested information:

.kibana       .kibana_1
sample-alias1 sample-index-1
sample-alias2 sample-index-2

Typically, for any operation you can find out what headers are available using the help parameter, and then use the h parameter to limit the output to only the headers that you care about.

Sort by a header

To sort the output by a header, use the s parameter:

GET _cat/{operation_name}?s={header_name_1},{header_name_2}

For example, to sort aliases by index and then by alias, send the following request:

GET /_cat/aliases?s=i,a
response = client.cat.aliases(
  params = { "s": "i,a" }
)

The response contains the requested information:

.kibana       .kibana_1      - - - -
sample-alias1 sample-index-1 - - - -
sample-alias2 sample-index-2 - - - -

Limit the number of rows

Introduced 3.10

To cap the number of rows returned, use the limit parameter:

GET _cat/{operation_name}?limit={number_of_rows}

The limit value must be a positive integer. A value of 0, a negative number, or a non-numeric value causes the request to fail with a 400 error. If you omit limit, the API returns all rows.

The limit parameter is applied after sorting (s), so combining the two returns the top N rows by the sort key. For example, the following request returns the five largest shards:

GET _cat/shards?s=store:desc&limit=5&v

The following request returns the 10 longest-running tasks:

GET _cat/tasks?s=running_time:desc&limit=10&v

If you do not specify s, the rows are returned in the default order of the operation, and limit returns the first N of those rows.

The limit parameter has the following behavior:

  • The parameter is validated for every CAT operation but is applied only by the non-paginated CAT APIs. The paginated List APIs (_list/*) ignore limit and use the size and next_token parameters instead because truncating a page would skip the rows between the limit and the page boundary.
  • The parameter reduces only the size of the response and the rendering work. OpenSearch still collects the data for all rows before applying the limit, so limit does not reduce the cost of building the table. To reduce that cost, restrict the request to specific indexes, for example, GET _cat/shards/my-index-*.
  • The limit parameter truncates the response and does not return an error. To reject requests that would return too many rows, use the cluster-level CAT response limit settings, which return a 429 error when the limit is exceeded.

Retrieve data in JSON format

By default, CAT APIs return data in text/plain format.

To retrieve data in JSON format, use the format=json parameter:

GET _cat/{operation_name}?format=json

For example, to retrieve aliases in JSON format, send the following request:

GET /_cat/aliases?format=json
response = client.cat.aliases(
  params = { "format": "json" }
)

The response contains data in JSON format:

[
  {"alias":".kibana","index":".kibana_1","filter":"-","routing.index":"-","routing.search":"-","is_write_index":"-"},
  {"alias":"sample-alias1","index":"sample-index-1","filter":"-","routing.index":"-","routing.search":"-","is_write_index":"-"},
  {"alias":"sample-alias2","index":"sample-index-2","filter":"-","routing.index":"-","routing.search":"-","is_write_index":"-"}
]

Other supported formats are YAML, CBOR, and Smile.

CAT API operations

The following CAT API operations are available.

Cluster and node information

Index and document information

Snapshot information

If you use the Security plugin, make sure you have the appropriate permissions.

⌂ 换网址