Semantic Conventions for Microsoft Azure Cosmos DB

Status: Experimental

The Semantic Conventions for Microsoft Cosmos DB extend and override the Database Semantic Conventions.

Attributes

db.system MUST be set to "cosmosdb" and SHOULD be provided at span creation time.

Cosmos DB instrumentation includes call-level (public API) surface spans and network spans. Depending on the connection mode (Gateway or Direct), network-level spans may also be created.

AttributeTypeDescriptionExamplesRequirement LevelStability
db.collection.namestringCosmos DB container name. [1]public.users; customersConditionally Required if availableExperimental
db.cosmosdb.connection_modestringCosmos client connection mode.gateway; directConditionally Required [2]Experimental
db.cosmosdb.consistency_levelstringAccount or request consistency level.Eventual; ConsistentPrefix; BoundedStaleness; Strong; SessionConditionally Required If available.Experimental
db.cosmosdb.regions_contactedstring[]List of regions contacted during operation in the order that they were contacted. If there is more than one region listed, it indicates that the operation was performed on multiple regions i.e. cross-regional call. [3]["North Central US", "Australia East", "Australia Southeast"]Conditionally Required If available.Experimental
db.cosmosdb.request_chargedoubleRequest units consumed for the operation.46.18; 1.0Conditionally Required when availableExperimental
db.cosmosdb.sub_status_codeintCosmos DB sub status code.1000; 1002Conditionally Required when response was received and contained sub-code.Experimental
db.namespacestringThe name of the database, fully qualified within the server address and port.customers; test.usersConditionally Required If available.Experimental
db.operation.namestringThe name of the operation or command being executed. [4]create_item; query_items; read_itemConditionally Required [5]Experimental
db.response.returned_rowsintCosmos DB row count in result set.10; 20Conditionally Required if response was received and returned any rowsExperimental
db.response.status_codestringCosmos DB status code. [6]200; 201Conditionally Required if response was receivedExperimental
error.typestringDescribes a class of error the operation ended with. [7]timeout; java.net.UnknownHostException; server_certificate_invalid; 500Conditionally Required If and only if the operation failed.Stable
server.portintServer port number. [8]80; 8080; 443Conditionally Required If not default (443).Stable
az.namespacestringAzure Resource Provider Namespace as recognized by the client. [9]Microsoft.DocumentDBRecommendedExperimental
db.cosmosdb.client_idstringUnique Cosmos client instance id.3ba4827d-4422-483f-b59f-85b74211c11dRecommendedExperimental
db.cosmosdb.request_content_lengthintRequest payload size in bytes.RecommendedExperimental
db.operation.batch.sizeintThe number of queries included in a batch operation. [10]2; 3; 4RecommendedExperimental
db.query.summarystringLow cardinality representation of a database query text. [11]SELECT wuser_table; INSERT shipping_details SELECT orders; get user by idRecommended [12]Experimental
db.query.textstringThe database query being executed. [13]SELECT * FROM wuser_table where username = ?; SET mykey ?Recommended [14]Experimental
server.addressstringName of the database host. [15]example.com; 10.1.2.80; /tmp/my.sockRecommendedStable
user_agent.originalstringFull user-agent string is generated by Cosmos DB SDK [16]cosmos-netstandard-sdk/3.23.0|3.23.1|1|X64|Linux 5.4.0-1098-azure 104 18|.NET Core 3.1.32|S|RecommendedStable
db.operation.parameter.<key>stringA database operation parameter, with <key> being the parameter name, and the attribute value being a string representation of the parameter value. [17]someval; 55Opt-InExperimental

[1] db.collection.name: It is RECOMMENDED to capture the value as provided by the application without attempting to do any case normalization.

[2] db.cosmosdb.connection_mode: if not gateway (the default value is assumed to be gateway).

[3] db.cosmosdb.regions_contacted: Region name matches the format of displayName in Azure Location API

[4] db.operation.name: The db.operation.name has the following list of well-known values. If one of them applies, then the respective value MUST be used.

Batch operations:

  • execute_batch

Bulk operations:

  • execute_bulk SHOULD be used on spans reported for methods like executeBulkOperations). which represents a bulk execution of multiple operations.
  • bulk_{operation name} (bulk_create_item, bulk_upsert_item, etc) SHOULD be used on spans describing individual operations (when they are reported) within the bulk. This pattern SHOULD be used when instrumentation creates span per each operation, but operations are buffered and then performed in bulk. For example, this applies when AllowBulkExecution property is configured on the Microsoft.Azure.Cosmos client.

Change feed operations:

  • query_change_feed

Conflicts operations:

  • delete_conflict
  • query_conflicts
  • read_all_conflicts
  • read_conflict

Container operations:

  • create_container
  • create_container_if_not_exists
  • delete_container
  • query_containers
  • read_all_containers
  • read_container
  • read_container_throughput
  • replace_container
  • replace_container_throughput

Database operations:

  • create_database
  • create_database_if_not_exists
  • delete_database
  • query_databases
  • read_all_databases
  • read_database
  • read_database_throughput
  • replace_database_throughput

Encryption key operations:

  • create_client_encryption_key
  • query_client_encryption_keys
  • read_all_client_encryption_keys
  • read_client_encryption_key
  • replace_client_encryption_key

Item operations:

  • create_item
  • delete_all_items_by_partition_key
  • delete_item
  • patch_item
  • query_items
  • read_all_items
  • read_all_items_of_logical_partition
  • read_many_items
  • read_item
  • replace_item
  • upsert_item

Permission operations:

  • create_permission
  • delete_permission
  • query_permissions
  • read_all_permissions
  • read_permission
  • replace_permission
  • upsert_permission

Stored procedure operations:

  • create_stored_procedure
  • delete_stored_procedure
  • execute_stored_procedure
  • query_stored_procedures
  • read_all_stored_procedures
  • read_stored_procedure
  • replace_stored_procedure

Trigger operations:

  • create_trigger
  • delete_trigger
  • query_triggers
  • read_all_triggers
  • read_trigger
  • replace_trigger

User operations:

  • create_user
  • delete_user
  • query_users
  • read_all_users
  • read_user
  • replace_user
  • upsert_user

User-defined function operations:

  • create_user_defined_function
  • delete_user_defined_function
  • query_user_defined_functions
  • read_all_user_defined_functions
  • read_user_defined_function

If none of them applies, it’s RECOMMENDED to use language-agnostic representation of client method name in snake_case. Instrumentations SHOULD document additional values when introducing new operations.

[5] db.operation.name: If readily available and if there is a single operation name that describes the database call. The operation name MAY be parsed from the query text, in which case it SHOULD be the single operation name found in the query.

[6] db.response.status_code: The status code returned by the database. Usually it represents an error code, but may also represent partial success, warning, or differentiate between various types of successful outcomes. Semantic conventions for individual database systems SHOULD document what db.response.status_code means in the context of that system. This attribute has stability level RELEASE CANDIDATE.

[7] error.type: The error.type SHOULD match the db.response.status_code returned by the database or the client library, or the canonical name of exception that occurred. When using canonical exception type name, instrumentation SHOULD do the best effort to report the most relevant type. For example, if the original exception is wrapped into a generic one, the original exception SHOULD be preferred. Instrumentations SHOULD document how error.type is populated.

[8] server.port: When observed from the client side, and when communicating through an intermediary, server.port SHOULD represent the server port behind any intermediaries, for example proxies, if it’s available.

[9] az.namespace: When az.namespace attribute is populated, it MUST be set to Microsoft.DocumentDB for all operations performed by Cosmos DB client.

[10] db.operation.batch.size: Operations are only considered batches when they contain two or more operations, and so db.operation.batch.size SHOULD never be 1. This attribute has stability level RELEASE CANDIDATE.

[11] db.query.summary: db.query.summary provides static summary of the query text. It describes a class of database queries and is useful as a grouping key, especially when analyzing telemetry for database calls involving complex queries. Summary may be available to the instrumentation through instrumentation hooks or other means. If it is not available, instrumentations that support query parsing SHOULD generate a summary following Generating query summary section. This attribute has stability level RELEASE CANDIDATE.

[12] db.query.summary: if readily available or if instrumentation supports query summarization.

[13] db.query.text: For sanitization see Sanitization of db.query.text. For batch operations, if the individual operations are known to have the same query text then that query text SHOULD be used, otherwise all of the individual query texts SHOULD be concatenated with separator ; or some other database system specific separator if more applicable. Even though parameterized query text can potentially have sensitive data, by using a parameterized query the user is giving a strong signal that any sensitive data will be passed as parameter values, and the benefit to observability of capturing the static part of the query text by default outweighs the risk. This attribute has stability level RELEASE CANDIDATE.

[14] db.query.text: Non-parameterized query text SHOULD NOT be collected by default unless there is sanitization that excludes sensitive data, e.g. by redacting all literal values present in the query text. See Sanitization of db.query.text. Parameterized query text SHOULD be collected by default (the query parameter values themselves are opt-in, see db.operation.parameter.<key>).

[15] server.address: When observed from the client side, and when communicating through an intermediary, server.address SHOULD represent the server address behind any intermediaries, for example proxies, if it’s available.

[16] user_agent.original: The user-agent value is generated by SDK which is a combination of
sdk_version : Current version of SDK. e.g. ‘cosmos-netstandard-sdk/3.23.0’
direct_pkg_version : Direct package version used by Cosmos DB SDK. e.g. ‘3.23.1’
number_of_client_instances : Number of cosmos client instances created by the application. e.g. ‘1’
type_of_machine_architecture : Machine architecture. e.g. ‘X64’
operating_system : Operating System. e.g. ‘Linux 5.4.0-1098-azure 104 18’
runtime_framework : Runtime Framework. e.g. ‘.NET Core 3.1.32’
failover_information : Generated key to determine if region failover enabled. Format Reg-{D (Disabled discovery)}-S(application region)|L(List of preferred regions)|N(None, user did not configure it). Default value is “NS”.

[17] db.operation.parameter: If a parameter has no name and instead is referenced only by index, then <key> SHOULD be the 0-based index. If db.query.text is also captured, then db.operation.parameter.<key> SHOULD match up with the parameterized placeholders present in db.query.text. This attribute has stability level RELEASE CANDIDATE.

The following attributes can be important for making sampling decisions and SHOULD be provided at span creation time (if provided at all):


db.cosmosdb.connection_mode has the following list of well-known values. If one of them applies, then the respective value MUST be used; otherwise, a custom value MAY be used.

ValueDescriptionStability
directDirect connection.Experimental
gatewayGateway (HTTP) connection.Experimental

db.cosmosdb.consistency_level has the following list of well-known values. If one of them applies, then the respective value MUST be used; otherwise, a custom value MAY be used.

ValueDescriptionStability
BoundedStalenessbounded_stalenessExperimental
ConsistentPrefixconsistent_prefixExperimental
EventualeventualExperimental
SessionsessionExperimental
StrongstrongExperimental

error.type has the following list of well-known values. If one of them applies, then the respective value MUST be used; otherwise, a custom value MAY be used.

ValueDescriptionStability
_OTHERA fallback error value to be used when the instrumentation doesn’t define a custom value.Stable

Example

KeyValue
Span name"read_item orders"
az.namespace"Microsoft.DocumentDB"
db.system"cosmosdb"
db.collection.name"orders"
db.namespace"ShopDb"
db.operation.name"read_item"
server.address"account.documents.azure.com"
db.cosmosdb.client_id"3ba4827d-4422-483f-b59f-85b74211c11d"
user_agent.original"cosmos-netstandard-sdk/3.23.0|3.23.1|1|X64|Linux 5.4.0-1098-azure 104 18|.NET Core 3.1.32|S|"
db.cosmosdb.request_content_length20
db.response.status_code201
db.cosmosdb.sub_status_code0
db.cosmosdb.request_charge7.43

Operation Level Metrics

The following metrics provide insights into Azure Cosmos DB client operation performance and behavior.

Metric: db.client.operation.duration

This metric is required.

It captures the total time taken by an Azure Cosmos DB operation. This metric follows the common db.client.operation.duration definition.

Refer db.client.cosmosdb.operation.request_charge metrics for dimensions.

Metric: db.client.response.returned_rows

This metric is required.

It captures the number of items returned by a query or feed operation in Azure Cosmos DB. It helps identify response sizes that may contribute to high latency, increased memory/CPU usage, or network call failures. This metric follows the common db.client.response.returned_rows definition.

Refer db.client.cosmosdb.operation.request_charge metrics for dimensions.

Metric: db.client.cosmosdb.operation.request_charge

This metric is required.

It captures the Request Units consumed by each operation in Azure Cosmos DB. Since Request Units serve as a form of throughput control within the Azure Cosmos DB database, monitoring their usage is crucial to avoid throttling.

this metric SHOULD be specified with ExplicitBucketBoundaries of [ 1, 5, 10, 25, 50, 100, 250, 500, 1000].

Explaining bucket configuration:

  1. 1, 5, 10: Low Usage Levels, These smaller buckets allow for precise tracking of operations that consume a minimal number of Request Units. This is important for lightweight operations, such as basic read requests or small queries, where resource utilization should be optimized. Monitoring these low usage levels can help ensure that the application is not inadvertently using more resources than necessary.
  2. 25, 50: Moderate Usage Levels, These ranges capture more moderate operations, which are typical in many applications. For example, queries that return a reasonable amount of data or perform standard CRUD operations may fall within these limits. Identifying usage patterns in these buckets can help detect efficiency issues in routine operations.
  3. 100, 250: Higher Usage Levels, These boundaries represent operations that may require significant resources, such as complex queries or larger transactions. Monitoring RUs in these ranges can help identify performance bottlenecks or costly queries that might lead to throttling.
  4. 500, 1000: Very High Usage Levels, These buckets capture operations that consume a substantial number of Request Units, which may indicate potentially expensive queries or batch processes. Understanding the frequency and patterns of such high RU usage can be critical in optimizing performance and ensuring the application remains within provisioned throughput limits.
NameInstrument TypeUnit (UCUM)DescriptionStability
db.client.cosmosdb.operation.request_chargeHistogram{request_unit}Request charge consumed by the operationExperimental
AttributeTypeDescriptionExamplesRequirement LevelStability
db.collection.namestringCosmos DB container name. [1]public.users; customersConditionally Required If available.Experimental
db.cosmosdb.consistency_levelstringAccount or request consistency level.Eventual; ConsistentPrefix; BoundedStaleness; Strong; SessionConditionally Required If available.Experimental
db.cosmosdb.sub_status_codeintCosmos DB sub status code.1000; 1002Conditionally Required when response was received and contained sub-code.Experimental
db.namespacestringThe name of the database, fully qualified within the server address and port.customers; test.usersConditionally Required If available.Experimental
db.operation.namestringThe name of the operation or command being executed. [2]findAndModify; HMSET; SELECTConditionally Required [3]Experimental
db.response.status_codestringDatabase response status code. [4]102; ORA-17002; 08P01; 404Conditionally Required [5]Experimental
error.typestringDescribes a class of error the operation ended with. [6]timeout; java.net.UnknownHostException; server_certificate_invalid; 500Conditionally Required If and only if the operation failed.Stable
server.portintServer port number. [7]80; 8080; 443Conditionally Required [8]Stable
db.cosmosdb.regions_contactedstring[]List of regions contacted during operation in the order that they were contacted. If there is more than one region listed, it indicates that the operation was performed on multiple regions i.e. cross-regional call. [9]["North Central US", "Australia East", "Australia Southeast"]Recommended If availableExperimental
server.addressstringName of the database host. [10]example.com; 10.1.2.80; /tmp/my.sockRecommendedStable

[1] db.collection.name: It is RECOMMENDED to capture the value as provided by the application without attempting to do any case normalization.

The collection name SHOULD NOT be extracted from db.query.text, unless the query format is known to only ever have a single collection name present.

For batch operations, if the individual operations are known to have the same collection name then that collection name SHOULD be used.

This attribute has stability level RELEASE CANDIDATE.

[2] db.operation.name: It is RECOMMENDED to capture the value as provided by the application without attempting to do any case normalization.

The operation name SHOULD NOT be extracted from db.query.text, unless the query format is known to only ever have a single operation name present.

For batch operations, if the individual operations are known to have the same operation name then that operation name SHOULD be used prepended by BATCH , otherwise db.operation.name SHOULD be BATCH or some other database system specific term if more applicable.

This attribute has stability level RELEASE CANDIDATE.

[3] db.operation.name: If readily available and if there is a single operation name that describes the database call. The operation name MAY be parsed from the query text, in which case it SHOULD be the single operation name found in the query.

[4] db.response.status_code: The status code returned by the database. Usually it represents an error code, but may also represent partial success, warning, or differentiate between various types of successful outcomes. Semantic conventions for individual database systems SHOULD document what db.response.status_code means in the context of that system. This attribute has stability level RELEASE CANDIDATE.

[5] db.response.status_code: If the operation failed and status code is available.

[6] error.type: The error.type SHOULD match the db.response.status_code returned by the database or the client library, or the canonical name of exception that occurred. When using canonical exception type name, instrumentation SHOULD do the best effort to report the most relevant type. For example, if the original exception is wrapped into a generic one, the original exception SHOULD be preferred. Instrumentations SHOULD document how error.type is populated.

[7] server.port: When observed from the client side, and when communicating through an intermediary, server.port SHOULD represent the server port behind any intermediaries, for example proxies, if it’s available.

[8] server.port: If using a port other than the default port for this DBMS and if server.address is set.

[9] db.cosmosdb.regions_contacted: Region name matches the format of displayName in Azure Location API

[10] server.address: When observed from the client side, and when communicating through an intermediary, server.address SHOULD represent the server address behind any intermediaries, for example proxies, if it’s available.


db.cosmosdb.consistency_level has the following list of well-known values. If one of them applies, then the respective value MUST be used; otherwise, a custom value MAY be used.

ValueDescriptionStability
BoundedStalenessbounded_stalenessExperimental
ConsistentPrefixconsistent_prefixExperimental
EventualeventualExperimental
SessionsessionExperimental
StrongstrongExperimental

error.type has the following list of well-known values. If one of them applies, then the respective value MUST be used; otherwise, a custom value MAY be used.

ValueDescriptionStability
_OTHERA fallback error value to be used when the instrumentation doesn’t define a custom value.Stable

Metric: db.client.cosmosdb.active_instance.count

This metric is required.

It captures the number of active instances at any given time. Best practices dictate that there should ideally be only one instance of the SDK client per Azure Cosmos DB account. Having multiple instances of the SDK client for the same account in a single process can lead to CPU or memory-related issues.

NameInstrument TypeUnit (UCUM)DescriptionStability
db.client.cosmosdb.active_instance.countUpDownCounter{instance}Number of active client instancesExperimental
AttributeTypeDescriptionExamplesRequirement LevelStability
server.portintServer port number. [1]80; 8080; 443Conditionally Required [2]Stable
server.addressstringName of the database host. [3]example.com; 10.1.2.80; /tmp/my.sockRecommendedStable

[1] server.port: When observed from the client side, and when communicating through an intermediary, server.port SHOULD represent the server port behind any intermediaries, for example proxies, if it’s available.

[2] server.port: If using a port other than the default port for this DBMS and if server.address is set.

[3] server.address: When observed from the client side, and when communicating through an intermediary, server.address SHOULD represent the server address behind any intermediaries, for example proxies, if it’s available.