Semantic conventions for database client calls

Status: Experimental

Span kind: MUST always be CLIENT.

The span name SHOULD be set to a low cardinality value representing the statement executed on the database. It MAY be a stored procedure name (without arguments), DB statement without variable arguments, operation name, etc. Since SQL statements may have very high cardinality even without arguments, SQL spans SHOULD be named the following way, unless the statement is known to be of low cardinality: <db.operation> <db.name>.<db.sql.table>, provided that db.operation and db.sql.table are available. If db.sql.table is not available due to its semantics, the span SHOULD be named <db.operation> <db.name>. It is not recommended to attempt any client-side parsing of db.statement just to get these properties, they should only be used if the library being instrumented already provides them. When it’s otherwise impossible to get any meaningful span name, db.name or the tech-specific database name MAY be used.

Connection-level attributes

These attributes will usually be the same for all operations performed over the same database connection. Some database systems may allow a connection to switch to a different db.user, for example, and other database systems may not even have the concept of a connection at all.

[1]: net.peer.name SHOULD NOT be set if capturing it would require an extra DNS lookup.

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

[3]: If different than inet and if any of net.sock.peer.addr or net.sock.host.addr are set. Consumers of telemetry SHOULD accept both IPv4 and IPv6 formats for the address in net.sock.peer.addr if net.sock.family is not set. This is to support instrumentations that follow previous versions of this document.

[4]: If defined for the address family and if different than net.peer.port and if net.sock.peer.addr is set.

[5]: If database type is in-process ("inproc"), recommended for other database types.

Additional attribute requirements: At least one of the following sets of attributes is required:

• net.peer.name
• net.sock.peer.addr

db.system 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.

Notes and well-known identifiers for db.system

The list above is a non-exhaustive list of well-known identifiers to be specified for db.system.

If a value defined in this list applies to the DBMS to which the request is sent, this value MUST be used. If no value defined in this list is suitable, a custom value MUST be provided. This custom value MUST be the name of the DBMS in lowercase and without a version number to stay consistent with existing identifiers.

It is encouraged to open a PR towards this specification to add missing values to the list, especially when instrumentations for those missing databases are written. This allows multiple instrumentations for the same database to be aligned and eases analyzing for backends.

The value other_sql is intended as a fallback and MUST only be used if the DBMS is known to be SQL-compliant but the concrete product is not known to the instrumentation. If the concrete DBMS is known to the instrumentation, its specific identifier MUST be used.

Back ends could, for example, use the provided identifier to determine the appropriate SQL dialect for parsing the db.statement.

When additional attributes are added that only apply to a specific DBMS, its identifier SHOULD be used as a namespace in the attribute key as for the attributes in the sections below.

Connection-level attributes for specific technologies

[1]: If setting a db.mssql.instance_name, net.peer.port is no longer required (but still recommended if non-standard).

Call-level attributes

These attributes may be different for each operation performed, even if the same connection is used for multiple operations. Usually only one db.name will be used per connection though.

[1]: In some SQL databases, the database name to be used is called “schema name”. In case there are multiple layers that could be considered for database name (e.g. Oracle instance name and schema name), the database name to be used is the more specific layer (e.g. Oracle schema name).

[2]: The value may be sanitized to exclude sensitive information.

[3]: If applicable and not explicitly disabled via instrumentation configuration.

[4]: When setting this to an SQL keyword, it is not recommended to attempt any client-side parsing of db.statement just to get this property, but it should be set if the operation name is provided by the library being instrumented. If the SQL statement has an ambiguous operation, or performs more than one operation, this value may be omitted.

For Redis, the value provided for db.statement SHOULD correspond to the syntax of the Redis CLI. If, for example, the HMSET command is invoked, "HMSET myhash field1 'Hello' field2 'World'" would be a suitable value for db.statement.

In CouchDB, db.operation should be set to the HTTP method + the target REST route according to the API reference documentation. For example, when retrieving a document, db.operation would be set to (literally, i.e., without replacing the placeholders with concrete values): GET /{db}/{docid}.

In Cassandra, db.name SHOULD be set to the keyspace name.

In HBase, db.name SHOULD be set to the HBase namespace.

Call-level attributes for specific technologies

[1]: It is not recommended to attempt any client-side parsing of db.statement just to get this property, but it should be set if it is provided by the library being instrumented. If the operation is acting upon an anonymous table, or more than one table, this value MUST NOT be set.

Cassandra

Separated for clarity.

[1]: This mirrors the db.sql.table attribute but references cassandra rather than sql. It is not recommended to attempt any client-side parsing of db.statement just to get this property, but it should be set if it is provided by the library being instrumented. If the operation is acting upon an anonymous table, or more than one table, this value MUST NOT be set.

Examples

MySQL

Redis

In this example, Redis is connected using a unix domain socket and therefore the connection string and net.sock.peer.addr are left out. Furthermore, db.name is not specified as there is no database name in Redis and db.redis.database_index is set instead.

MongoDB