Service limits for ES Serverless applications-Elasticsearch(ES)-阿里云帮助中心

To ensure that your services run smoothly, Alibaba Cloud Elasticsearch Serverless (ES Serverless) sets limits on the resources and usage of each application. If a task fails due to insufficient quotas or service limits, you can request to adjust them.

Service limits

Service limits are system-level restrictions defined within the Serverless platform. They apply to the management of indexes, shards, and data, and to read and write requests within an application. If a limit is exceeded, the Serverless service rejects the relevant request or applies fencing to the application.

Note

The service limits listed in this topic are the default limits for new applications. For existing applications, the limits that are displayed in the console prevail.
You can request to modify the service limits. For more information, see Request to modify application quotas and service limits.
For descriptions of the basic concepts related to service limits, see Terms.

Limit item	Limitations	Default limit
index.mapping.depth.limit	Maximum depth of a nested JSON object.	[1 - 20]
index.mapping.field_name_length.limit	Maximum length of a field name.	[1 - 100]
index.mapping.total_fields.limit	Total number of single-indexed fields	[1 - 1000]
index.max_adjacency_matrix_filters	Maximum number of adjacency matrix filters in an index.	[0 - 100]
index.max_docvalue_fields_search	Maximum number of docvalue_fields.	[1 - 100]
index.max_inner_result_window	Maximum number of results for an inner query.	[1 - 100]
index.max_ngram_diff	Maximum n-gram distance	[0 - 1]
index.max_refresh_listeners	Maximum concurrent waits	[0 - 20]
index.max_regex_length	Maximum length of a regular expression.	[0 - 50]
index.max_rescore_window	Maximum number of results for rescoring.	[1 - 10000]
index.max_result_window	Maximum number of query results.	[1 - 10000]
index.max_script_fields	Maximum number of script_fields.	[1 - 32]
index.max_shingle_diff	Maximum Shingle Distance	[0 - 3]
index.max_terms_count	Maximum number of terms in a single query.	[0 - 1024]
index_quota	Maximum number of indexes.	500
max_cu_per_search	Maximum number of CUs consumed by a single query.	600
max_cu_per_shard	Maximum number of CUs consumed by a single shard in a single query.	10
max_doc_size	Maximum size of a single document (MB).	1
max_index_metadata_update_qps	Maximum QPS for index metadata updates (requests per 30 seconds).	1000
max_pipeline_num	Maximum number of pipelines.	100
max_prefix_length	Maximum length of the prefix string in a prefix query.	[0 - 50]
max_refresh_qps	Maximum refresh rate (requests per second).	5
max_size_per_request	Maximum size of a single request (MB).	30
max_storage_per_cu	Maximum storage-to-CU ratio. Increasing this ratio can affect query and write performance. Adjust this value with caution.	1600
max_timeout	Maximum timeout duration (seconds).	300
max_timeout_per_request	Maximum duration of a single query (seconds).	300
max_wildcard_length	Maximum length of the wildcard string in a wildcard query.	[0 - 50]
max_write_size_per_request	Maximum size of a single write request (MB).	5
search.max_buckets	Maximum number of buckets for a single aggregation.	10000
search.max_keep_alive	Maximum keep-alive time for a search (seconds).	900

Request to modify service limits

Log on to the Elasticsearch Serverless console.
In the top menu bar, select the region where your application is located.
In the navigation pane on the left, click Application Management.
In the application list, click the name of your application.
In the left navigation pane, choose Application Management > Service Quota.
Request to modify the service limits.
1. On the Quota Overview tab, click Modify Quota.
2. For the target limit item, click the icon in the Running Value column. In the dialog box that appears, enter the target value and click Confirm.
3. After modifying the target limits, click Submit Changes at the top of the page.
  Note
  Modifying service limits requires manual approval and is typically completed within one business day. For urgent requests, you can submit a ticket or join the Elasticsearch Serverless user group on DingTalk (group ID: 11205017670 or 72335013004) for assistance.
(Optional) Click the Request History tab to view the history, status, and values of your requests.
- In the Request Status column, you can view the status of your request.
- In the Requested Item column, click a Requested Item to view its historical, requested, and effective values.
  Note
  - The effective value may differ from the requested value. Check the effective value promptly after your request is approved.
  - Your request may be rejected if the requested value is unreasonable or could negatively impact your service. If your request is rejected, you can try other solutions or submit a ticket for assistance.
- To adjust a pending request, click Revoke in the Actions column and submit a new request.
  Note
  You cannot modify application quotas or service limits again until the current approval process is complete.

Terms

Concepts	Description
Application	An application is the basic management unit in Elasticsearch Serverless. It is used for resource isolation and access control. Each application has a unique endpoint, username, and password. You can use different applications to manage and store data for different services or projects.
Index	An index is a collection of documents with a similar structure. It is a logical namespace that points to one or more shards and is used to separate different datasets within the same service.
Document	A document is the smallest unit of information in Elasticsearch data storage that can be indexed and retrieved.
Shard	Elasticsearch is a distributed search engine that stores a complete index across multiple shards.
Replica	Both primary shards and replica shards can handle query requests, but only primary shards can handle indexing requests. Replicas provide fault recovery if a primary shard is corrupted or lost. They also improve query efficiency.