[Monitor-Opentelemetry] Expose a way to disable metrics (while keeping instrumentation) in instrumentation options

[Apokalypt]

Is your feature request related to a problem? Please describe.
Right now, when we are enabling instrumentation of a SDK through automatic configuration in useAzureMonitor we only have one property to enable/disable instrumentation and metrics. However, in some cases (like ours) we do not need metrics and don't want them as they are generating useless cost

Describe the solution you'd like
In the instrumentation options when calling useAzureMonitor of @azure/monitor-opentelemetry, I would like to have a property like disableMetrics to enable instrumentations but not register metrics, for example :

Describe alternatives you've considered

• Disable instrumentation in useAzureMonitor and then register manually instrumentation for the library

Additional context
N/A

[JacksonWeber]

@Apokalypt between the solution mentioned here: open-telemetry/opentelemetry-js#4487 and disabling the metrics API, it doesn't appear necessary to provide explicit config for this use-case. Please let me know if this doesn't cover your scenario.

[Apokalypt]

@Apokalypt between the solution mentioned here: open-telemetry/opentelemetry-js#4487 and disabling the metrics API, it doesn't appear necessary to provide explicit config for this use-case. Please let me know if this doesn't cover your scenario.

The solutions provided implies to have access to the SDK instantiation which is not the case when using @azure/monitor-opentelemetry . So, in this case, there is no way to provide a view with an aggregation "Aggregation.drop()" to drop metrics of a specific libraries
Disabling the metrics API completely doesn't seem to work either. Even with that, the instrumentation metrics are received by our service.

Here is a code that you can use to test and reproduce :

const { useAzureMonitor } = require('@azure/monitor-opentelemetry')
const { Resource } = require('@opentelemetry/resources')
const { ATTR_SERVICE_INSTANCE_ID } = require('@opentelemetry/semantic-conventions/incubating')
const { ATTR_SERVICE_NAME, ATTR_SERVICE_VERSION } = require('@opentelemetry/semantic-conventions')

useAzureMonitor({
    azureMonitorExporterOptions: { key: process.env.APPLICATIONINSIGHTS_CONNECTION_STRING },
    resource: new Resource({
        [ATTR_SERVICE_NAME]: 'test-metrics-express',
        [ATTR_SERVICE_VERSION]: '1.0.0',
        [ATTR_SERVICE_INSTANCE_ID]: process.env.WEBSITE_INSTANCE_ID
    }),

    instrumentationOptions: {
        http: { enabled: true },

        azureSdk: { enabled: false },
        mongoDb: { enabled: false },
        redis4: { enabled: false },
        mySql: { enabled: false },
        redis: { enabled: false },
        bunyan: { enabled: false },
        winston: { enabled: false },
        postgreSql: { enabled: false }
    },
    browserSdkLoaderOptions: { enabled: false },

    samplingRatio: 1,

    enableLiveMetrics: false,
    enableStandardMetrics: false,
    enableTraceBasedSamplingForLogs: true
})

const { metrics } = require('@opentelemetry/api')
metrics.disable()

const express = require('express')

const app = express()
    .disable('x-powered-by')

app.get('/', (req, res) => {
    res.status(200).send({ })
})

app.listen(3000, () => {
    console.debug('Server started on port 3000')
})

And in the table customMetrics we can see :