[Proposal] Semantic Conventions for IBM Informix DB (Performance Metrics)

Propose new conventions

Proposal:

Introduce New Semantic Conventions for Informix Database in OpenTelemetry

Summary

This proposal introduces semantic conventions for monitoring IBM Informix Database within the OpenTelemetry project.
Informix is a widely used database system known for its efficiency in OLTP (Online Transaction Processing) and time-series data handling. Currently, OpenTelemetry lacks specific semantic attributes for Informix. This proposal aims to standardize the metrics for Informix to ensure better observability, performance monitoring, and interoperability across telemetry data sources.

Proposed Semantic Conventions (Specific to Informix DB)

Performance Metrics

    1. Cache hit ratio
    2. Lock Count
    3. Lock Time
    4. Sequential Scan Count
    5. Sequential Table Count
    6. Overflow Transaction Count
    7. Overflow User Count
    8. Lock waits
    9. Cache Read Ratio
    10. Cache Write Ratio
    11. LRU writes

Performance Metrics

Metric: db.cache.hit

This metric is recommended.

Name	Instrument Type	Units (UCUM)	Description
db.cache.hit	Gauge	1	The cache hit ratio/percentage

Attribute	Type	Description	Example	Requirement Level
type	string	The type of cache.	NORMAL;RECYCLE;FAST;KEEP	Recommended

Metric: db.lock.count

This metric is recommended.

Name	Instrument Type	Units (UCUM)	Description
db.lock.count	UpDownCounter	{lock}	The number of database locks.

Attribute	Type	Description	Example	Requirement Level
type	string	The type of lock.	Row-level Lock	Recommended

Metric: db.lock.time

This metric is recommended.

Name	Instrument Type	Units (UCUM)	Description
db.lock.time	UpDownCounter	s	The lock elapsed time.

Attribute	Type	Description	Example	Requirement Level
lock_id	string	The lock ID.	0x987654321	Required
blocking_sess_id	string	The blocking session identifier.	28871001	Recommended
blocker_sess_id	string	The blocker session identifier.	28871041	Recommended
locked_obj_name	string	The locked object name.	gv$instance	Recommended

Metric: db.seq.scan.count

This metric is optional.

Name	Instrument Type	Units (UCUM)	Description
db.seq.scan.count	UpDownCounter	{scan}	Count of Sequential Scans for a table.

Attribute	Type	Description	Example	Requirement Level
table_name	string	Name of the table.	user ; product	Required

Metric: db.seq.scan.table.count

This metric is optional.

Name	Instrument Type	Units (UCUM)	Description
db.seq.scan.table.count	UpDownCounter	{tableScan}	Number of tables which is having sequential scan.

Metric: db.overflow.lock.count

This metric is optional.

Name	Instrument Type	Units (UCUM)	Description
db.overflow.lock.count	UpDownCounter	{overflowLock}	Number of times a thread attempted to acquire a lock when no locks were available.

Metric: db.overflow.transaction.count

This metric is optional.

Name	Instrument Type	Units (UCUM)	Description
db.overflow.transaction.count	UpDownCounter	{overflowTransaction}	Number of times a thread attempted to allocate an entry in the transaction table when no entries in the shared-memory table were available.

Metric: db.overflow.user.count

This metric is optional.

Name	Instrument Type	Units (UCUM)	Description
db.overflow.user.count	UpDownCounter	{overflowUser}	Number of times a User thread attempted to acquire a lock when no locks were available.

Metric: db.lock.waits

This metric is optional.

Name	Instrument Type	Units (UCUM)	Description
db.lock.waits	UpDownCounter	{lockWaits}	Number of threads waiting for lock.

Metric: db.cache.read.ratio

This metric is optional.

Name	Instrument Type	Units (UCUM)	Description
db.cache.read.ratio	Gauge	1	Percentage of page reads for this buffer pool that were satisfied by a cached page image

Metric: db.cache.write.ratio

This metric is optional.

Name	Instrument Type	Units (UCUM)	Description
db.cache.write.ratio	Gauge	1	Percentage of page Writes for this buffer pool that were satisfied by a cached page image

Metric: db.lru.writes

This metric is optional.

Name	Instrument Type	Units (UCUM)	Description
db.lru.writes	UpDownCounter	{lruWrites}	Number of Least Recently Used Writes.

Notes:

• The database server performs LRU (Least Recently Used) writes as background writes that typically occur when the percentage of dirty buffers (pages that are not accessed) exceeds the percent that is specified for lru_max_dirty in the BUFFERPOOL configuration parameter.

Start with the proposal's metric tables and the linked metric-requirement guidance. Review how this repository represents database semantic conventions before assessing the Informix names, instruments, units, and attributes. Done means the proposal has an accepted, consistent set of conventions for the listed Informix performance metrics.