Database metrics - adding conventions for database server metrics

[niuxiaojuan-github]

Hi, We have following proposal for database server metrics, would like to hear some feedback from community.

We referenced to General Guidelines, Attribute Naming, Requirement Levels and current database metrics for connection pool, system metrics, process metric to draft this.

We attempted to extract metrics that are as generic as possible, but this is not an easy task due to the diversity of the databases, we are still working on it, and would like to get some guidance and feedback based on currently what we have.

Resource attributes

Database

Description: A database.

Attribute Key	Type	Description	Example	Requirement Level
db.system	string	An identifier for the database management system (DBMS) product being used.	db2, damengdb	Required
db.version	string	The version of the database	V11.5, V8	Required
server.address	string	Name of the database host.	db.testdb.com	Recommended
server.port	int	database listen port	50000, 5236	Recommended
db.name	string	This attribute is used to report the name of the database being accessed. For commands that switch the database, this should be set to the target database	BLUDB, DAMENG	Conditionally Required: If applicable.

Instance

Description: A database instance. Conditionally Required.

Attribute Key	Type	Description	Example	Requirement Level
db.instance.name	string	The db instance name	db2inst1, DMSERVER	Conditionally Required: If applicable.
db.instance.number	int	The db instance number	1	Conditionally Required
db.instance.host	string	The hostname of the db instance	db.testdb.com	Conditionally Required

Common metric instruments

All metrics in db. instruments should be attached to a Database resource and therefore inherit its attributes, like db.system.

All metrics in db.instance instruments should be attached to a Instance resource and therefore inherit its attributes, like db.instance.name.

Availability

Name	Description	Units	Instrument Type	Value Type	Attribute Key(s)	Attribute Values
db.status	Recommended. The status of the database. 1 (Active), 0 (Inactive)	{status}	Gauge	Int	state (Required)	Active, Inactive
db.instance.status	Recommended. The status of the db instance. 1 (Active), 0 (Inactive)	{status}	Gauge	Int	state (Required)	Active, Inactive
db.instances.count	Recommended. The total number of db instances.	{instance}	UpDownCounter	Int
db.instances.active.count	Recommended. The total number of active db instances.	{instance}	UpDownCounter	Int

Throughput

Name	Description	Unit	Instrument Type	Value Type	Attribute Key(s)	Attribute Values
db.sessions.count	Recommended. The number of database sessions.	{session}	UpDownCounter	Int
db.sessions.active.count	Recommended. The number of active database sessions.	{session}	UpDownCounter	Int
db.transactions.count	Recommended. The number of completed transactions.	{transaction}	UpDownCounter	Int
db.transactions.rate	Recommended. The number of transactions per second.	{transaction}	Gauge	Double
db.transactions.latency	Recommended. The average transaction latency.	s	Gauge	Double
db.sql.count	Recommended. The number of SQLs	{sql}	UpDownCounter	Int64
db.sql.rate	Recommended. The number of SQL per second.	{sql}	Gauge	Double
db.sql.latency	Recommended. The average SQL latency.	s	Gauge	Double
db.io.read.rate	Recommended. The physical read per second.	By	Gauge	Double
db.io.write.rate	Recommended. The physical write per second.	By	Gauge	Double

Performance

Name	Description	Unit	Instrument Type	Value Type	Attribute Key(s)	Attribute values
db.cache.hit	The cache hit ratio/percentage	1	UpDownCount	Double	type	(cache type)

[niuxiaojuan-github]

Discussed on Sept 11 meeting, @jsuereth suggested creating a project proposal for a semconv working group on databases.

[XiangSu-IBM]

@niuxiaojuan-github Hi Xiaojuan, Sorry for let you wait. I just checked the list above and I think this list is very great and most index are common to all database products. But just talk about DB2, here we have a product called DMC(data management console). It is a is an integrated database management tools platform that you can use to manage DB2 LUW. The metrix data collected by DMC is very useful for DB2 user to monitor the running of DB2. Perhaps the metrix of DMC can help something. Please refer to below link about DMC collected DB2 related metrix. https://www.ibm.com/docs/en/db2-data-mgr-console/3.1.x?topic=metrics-summary