Why is PostgreSQL peer.address a JDBC string?

[pauldraper]

jdbc:postgresql://host:5432/database

Why? Do tools normally use JDBC URLs as an address?

All tools I know of, including psql, use the postgresql (or postgres) scheme.

psql postgresql://host:5432/database

[vmarchaud]

I believe the spec require to use a common (eg: jdbc) format: https://github.com/open-telemetry/opentelemetry-specification/blob/main/specification/trace/semantic_conventions/database.md#connection-level-attributes-for-specific-technologies

[pauldraper]

In my reading, I don't see anything about peer.address (or net.peer.address).

[github-actions]

This issue is stale because it has been open 60 days with no activity. Remove stale label or comment or this will be closed in 14 days.

[github-actions]

This issue was closed because it has been stale for 14 days with no activity.

[pauldraper]

I still have this question.

I don't see anything about peer.address in the standard.

[dyladan]

It appears it is just drastically outdated. The spec at the time of writing the postgres instrumentation was this: https://github.com/open-telemetry/opentelemetry-specification/blob/c8109895a989480028b282fa796ed87ad1421b8a/specification/data-database.md which specified the JDBC substring.

The recent spec is quite different and requires:

• db.system - postgresql
• One of net.peer.name OR net.peer.ip - hostname or ip respectively
• net.peer.port only if using nonstandard port
• net.transport recommended in general, required only if using in-process - tcp or inproc

db.connection_string is not required but seems to be the replacement for the jdbc string.

This most likely also affects the mysql and mysql2 instrumentations

[dyladan]

Created #781 to address this