Enhance metric name disambiguation using attributes #41303
Description
Component(s)
exporter/bmchelix
Is your feature request related to a problem? Please describe.
This update improves the uniqueness and clarity of exported metric names by enriching them based on label attributes that vary across data points sharing the same entityId and metricName.
Currently, to ensure that each metric name is uniquely identified in BMC Helix Operations Management (BHOM), users use OTTL to append the value of a distinguishing attribute such as state to the metric name.
Example:
- context: datapoint
statements:
- set(metric.name, Concat([metric.name, attributes["cpu.mode"]], ".")) where attributes["cpu.mode"] != nilThis transformation will rename metrics like system.cpu.time to system.cpu.time.user or system.cpu.time.system, depending on the cpu.mode value.
So modifying metric.name within a datapoint context causes a conflict when multiple datapoints belong to the same metric, as the name change applies to the entire metric rather than individual datapoints.
Describe the solution you'd like
To ensures the uniqueness of metric names within BMC Helix, we need to enrich metric names with datapoint attributes in the helix exporter:
Example transformation:
Input:
"metrics": [
{
"name": "system.cpu.time",
"unit": "s",
"description": "Seconds each logical CPU spent on each mode",
"gauge": {
"dataPoints": [
{
"asDouble": 10,
"timeUnixNano": "1544712660300000000",
"attributes": [
{
"key": "cpu.mode",
"value": {
"stringValue": "system"
}
}
]
},
{
"asDouble": 20,
"timeUnixNano": "1544712660300000000",
"attributes": [
{
"key": "cpu.mode",
"value": {
"stringValue": "user"
}
}
]
}
]
}
}
]Helix Output:
{
"MetricName": "system.cpu.time.user",
"Samples": [
{
"Value": 10,
"Timestamp": 1544712660300000000
}
]
},
{
"MetricName": "system.cpu.time.user",
"Samples": [
{
"Value": 20,
"Timestamp": 1544712660300000000
}
]
}Describe alternatives you've considered
OTTL!
But works only if there is one metric per state.
Additional context
No response
Tip
React with 👍 to help prioritize this issue. Please use comments to provide useful context, avoiding +1 or me too, to help us triage it. Learn more here.