metrics20 / spec Goto Github PK

I'm toying with the idea of supporting passing in metrics "adapters" to a library and then internal to the library just calling said adapters with a standardized metrics format.

Example

// pass in prometheus adapter
const instance = new Library({ metrics: prometheusAdapter })

// inside the library, use the passed in adapter to send out metrics.
prometheusAdapter({
    what: 'response_time',
    http_code: 206,
    http_method: 'GET',
    host: 'dfsproxy1',
    service: 'proxy-server'
    stat: 'upper_90',
    swift_type: 'object'
    target_type: 'gauge'
    unit: 'ms'
})

Do you think metrics2.0 would be a good choice for the wire format? It should have most things needed in order to write adapters for the various metrics systems like prometheus etc. right?

Since tagging formats and systems tend to have alot of "pass through" context (much like HTTP headers). "Unique-ness" or IDs should not be based on the tag ordering, as down stream systems can continue to add tags. And like HTTP headers the "value" is what is important (labels are important to, but their ordering is not). So i propose that any "unique-id" generator needs to sort the tags by name.

In this fashion

name1=val1 name2=val2 and name2=val2 name1=val1 really mean the same thing and thus should not be considered "different"

i.e.

input of

nameZ=val1 nameY=val2 would be converted whenever the ID was generated to makeMeYourStorageId("nameY=val2 nameZ=val1")

What population comprises your target audience for the metrics20.org website and the spec called "Metrics 2.0"?

What would be a good tag name to specify data belonging to a specific tenant in a multi tenant system? At raintank we are almost completely compatible with the spec, we just need to formalize what tag to use to identify tenants. We use the org[anisation] terminology so we currently use org_id , does anyone know of commenly used alternatives ? tenant_id ? tid ?

Please consider not innovating with yet another textual/syntactic format, but instead adopt an existing one like JSON or YAML.

"status" seems to be a clearer term for whether memory is free/in use etc

There is a lot of the Metrics 2.0 implemented in InfluxDB / Prometheus (adopted by the CNCF for Cloud Native deployments).. yet I see no reference to those engines on the metrics20.org website ?

Do they really not match the model here, or is it an overlook ?