Explicit histogram boundaries do not match the otel spec

[artdent]

What happened?

Steps to Reproduce

Initialize sdk-metrics and create a histogram with the default settings. Observe the chosen histogram buckets.

Expected Result

The chosen boundaries should match the metrics spec:

The Default Value represents the following buckets (heavily influenced by the default buckets of Prometheus clients, e.g. Java and Go):
(-∞, 0], (0, 5.0], (5.0, 10.0], (10.0, 25.0], (25.0, 50.0], (50.0, 75.0], (75.0, 100.0], (100.0, 250.0], (250.0, 500.0], (500.0, 750.0], (750.0, 1000.0], (1000.0, 2500.0], (2500.0, 5000.0], (5000.0, 7500.0], (7500.0, 10000.0], (10000.0, +∞). SDKs SHOULD use the default value when boundaries are not explicitly provided, unless they have good reasons to use something different (e.g. for backward compatibility reasons in a stable SDK release).

Specifically, the default boundaries should be [0, 5, 10, 25, 50, 75, 100, 250, 500, 750, 1000, 2500, 5000, 7500, 10000].

Actual Result

The default boundaries in the node metrics sdk instead are [0, 5, 10, 25, 50, 75, 100, 250, 500, 750, 1000], which is not as useful for e.g. bucketing slow RPCs.

Additional Details

It looks like the defaults are configured here in the sdk-metrics package.

OpenTelemetry Setup Code

No response

package.json

{
  "dependencies": {
    "@opentelemetry/api": "1.3.0",
    "@opentelemetry/sdk-node": "0.34.0",
    "@opentelemetry/sdk-trace-base": "1.8.0",
  }
}

Relevant log output

No response

[pichlermarc]

Thanks for reaching out, looks like this spec change flew under the radar for us. 😓

AFAIK adding these buckets is not a breaking change, and other SDKs made this change after GA as well, so adding these should be safe. 🙂

[Bharath-Ganesh]

@pichlermarc Can I take this up? I'm reading the documentation to get a better understanding.
I'll revert back to you with my findings.

[pichlermarc]

@pichlermarc Can I take this up? I'm reading the documentation to get a better understanding. I'll revert back to you with my findings.

It's yours. 🎉 Thanks for picking this up. 🙂
If you have any questions, feel free to reach out here or on the CNCF Slack 🙂

[Bharath-Ganesh]

Great, I'll do that.
By using the default settings, I'm attempting to make a histogram.
https://opentelemetry.io/docs/reference/specification/metrics/api/#histogram-creation

While attempting to execute the npm test, I encountered a problem.
If I can't fix the problem below, I'll consult with the Slack team.

[pichlermarc]

@Bharath-Ganesh were you able to resolve the issue? 🤔
I did spin up a Windows VM and installed Node 14 and it looks like there seem to be some issues indeed, however I could not reproduce the exact problem from your screenshot.

Could you try if the same problem persists with Node 16? 🤔 Maybe there's a problem with some of our tooling.

[Bharath-Ganesh]

I tried to spin up a Ubuntu VM to run the test, and it's still failing.
node version: 18

[pichlermarc]

Interesting, could you include the actual error message in the logs? 🤔
With node 18, you'll need to run the tests using NODE_OPTIONS=--openssl-legacy-provider as we're using an older version of webpack that needs this flag.

[ashutosh887]

I would like to work on this @pichlermarc @artdent

[pichlermarc]

@ashutosh887 I wonder if @Bharath-Ganesh is still working on this, it may be worth syncing up with them to avoid some duplicate work. Feel free to open a PR then. 🙂

[Bharath-Ganesh]

@ashutosh887 I'm working on this. I got caught up with some work, sorry for the delay.

[ashutosh887]

Thanks @pichlermarc
Sure @Bharath-Ganesh let me know if you need any help in this

[ashutosh887]

Hey @Bharath-Ganesh are you still working on this?

cc: @pichlermarc

[Bharath-Ganesh]

@ashutosh887 I'm pretty swamped. You could take this up.

[ashutosh887]

Thanks @Bharath-Ganesh
@pichlermarc Please assign this issue to me