OTEL - Metrics

[helio-frota]

Jens wrote in the chat:

• ..., and we do lack metrics right now. at least, compared to v1.

Looking at v1 metrics, I know we can discard a), b) and d) ( no kafka, no tantivy, and we already have default actix-web metrics )

c) is not clear to me, but if we discard that, then we end up with no manual/custom metrics and that is a good thing for version 2x because we can have default-metrics soon and think about custom/metrics later.

ATM I can see this in case we discard a,b,c,d above:

Pros

• Without custom metrics we can have something running sooner than later
• We can remove prometheus and actix-web-prom since actix-web-opentelemetry already have support for metrics.
• We end up with the same approach used in traces app -> OTEL Collector --> Vendor back-end

Cons

• Need to figure out how to add/have custom metrics with otel* crates (and continue with app -> OTEL Collector --> Vendor back-end instead of app -> prometheus)

• Without custom/metrics we end up with http related metrics only

  • example:

Metrics
Resource
	 ->  service.name=String(Owned("m-o"))
	Instrumentation Scope #0
		Name         : actix-web-opentelemetry
		Version  : "0.20.1"
		SchemaUrl: "https://opentelemetry.io/schemas/1.28.0"
Metric #0
		Name         : http.server.duration
		Description  : Measures the duration of inbound HTTP requests.
		Unit         : s
		Type         : Histogram
		Temporality  : Cumulative
		Histogram DataPoints
		DataPoint #0
			StartTime    : 2025-01-24 14:07:10.140051
			EndTime      : 2025-01-24 14:08:10.140579
			Count        : 11
			Sum          : 19.837263478000004
			Min          : 1.802292795
			Max          : 1.804464909
			Attributes   :
				 ->  http.route: /
				 ->  http.request.method: GET
				 ->  network.protocol.version: 1.1
				 ->  server.address: localhost
				 ->  server.port: 8080
				 ->  url.scheme: http
				 ->  http.response.status_code: 200
Metric #1
		Name         : http.server.active_requests
		Description  : Measures the number of concurrent HTTP requests that are currently in-flight.
		Unit         :
		Type         : Sum
		Sum DataPoints
		Monotonic    : false
		Temporality  : Cumulative
		DataPoint #0
			StartTime    : 2025-01-24 14:07:10.140114
			EndTime      : 2025-01-24 14:08:10.140644
			Value        : 0
			Attributes   :
				 ->  http.request.method: GET
				 ->  http.route: /
				 ->  network.protocol.version: 1.1
				 ->  server.address: localhost
				 ->  server.port: 8080
				 ->  url.scheme: http
Metric #2
		Name         : http.server.request.size
		Description  : Measures the size of HTTP request messages (compressed).
		Unit         : By
		Type         : Histogram
		Temporality  : Cumulative
		Histogram DataPoints
		DataPoint #0
			StartTime    : 2025-01-24 14:07:10.140175
			EndTime      : 2025-01-24 14:08:10.140685
			Count        : 11
			Sum          : 0
			Min          : 0
			Max          : 0
			Attributes   :
				 ->  http.request.method: GET
				 ->  http.route: /
				 ->  network.protocol.version: 1.1
				 ->  server.address: localhost
				 ->  server.port: 8080
				 ->  url.scheme: http
Metric #3
		Name         : http.server.response.size
		Description  : Measures the size of HTTP response messages (compressed).
		Unit         : By
		Type         : Histogram
		Temporality  : Cumulative
		Histogram DataPoints
		DataPoint #0
			StartTime    : 2025-01-24 14:07:10.140219
			EndTime      : 2025-01-24 14:08:10.140719
			Count        : 11
			Sum          : 55
			Min          : 5
			Max          : 5
			Attributes   :
				 ->  http.route: /
				 ->  http.request.method: GET
				 ->  network.protocol.version: 1.1
				 ->  server.address: localhost
				 ->  server.port: 8080
				 ->  url.scheme: http
				 ->  http.response.status_code: 200

[helio-frota]

Related with #1198

[helio-frota]

Custom metrics example without prometheus and actix-web-prom crates screenshot

[helio-frota]

This can be closed 👍