feat: async call otel metrics

[deniseli]

These are just the basic metrics. Queue depth up-down counters should come next

ScopeMetrics #3
ScopeMetrics SchemaURL: 
InstrumentationScope ftl.async_call 

Metric #0
Descriptor:
     -> Name: ftl.async_call.acquired
     -> Description: the number of times that the controller tries acquiring an async call
     -> Unit: 1
     -> DataType: Sum
     -> IsMonotonic: true
     -> AggregationTemporality: Cumulative

NumberDataPoints #0
Data point attributes:
     -> ftl.async_call.origin: Str(sub:echo.sub)
     -> ftl.async_call.time_since_scheduled_at_ms: Int(24)
     -> ftl.async_call.verb.ref: Str(echo.echoSinkOne)
     -> ftl.module.name: Str(echo)
     -> ftl.status.succeeded: Bool(true)
StartTimestamp: 2024-07-31 00:33:53.897947 +0000 UTC
Timestamp: 2024-07-31 00:34:08.898478 +0000 UTC
Value: 1

Metric #1
Descriptor:
     -> Name: ftl.async_call.executed
     -> Description: the number of times that the controller tries executing an async call
     -> Unit: 1
     -> DataType: Sum
     -> IsMonotonic: true
     -> AggregationTemporality: Cumulative

NumberDataPoints #0
Data point attributes:
     -> ftl.async_call.origin: Str(sub:echo.sub)
     -> ftl.async_call.time_since_scheduled_at_ms: Int(41)
     -> ftl.async_call.verb.ref: Str(echo.echoSinkOne)
     -> ftl.module.name: Str(echo)
     -> ftl.status.succeeded: Bool(true)
StartTimestamp: 2024-07-31 00:33:53.897952 +0000 UTC
Timestamp: 2024-07-31 00:34:08.898485 +0000 UTC
Value: 1

Metric #2
Descriptor:
     -> Name: ftl.async_call.completed
     -> Description: the number of times that the controller tries completing an async call
     -> Unit: 1
     -> DataType: Sum
     -> IsMonotonic: true
     -> AggregationTemporality: Cumulative

NumberDataPoints #0
Data point attributes:
     -> ftl.async_call.origin: Str(sub:echo.sub)
     -> ftl.async_call.time_since_scheduled_at_ms: Int(46)
     -> ftl.async_call.verb.ref: Str(echo.echoSinkOne)
     -> ftl.module.name: Str(echo)
     -> ftl.status.succeeded: Bool(true)
StartTimestamp: 2024-07-31 00:33:53.897956 +0000 UTC
Timestamp: 2024-07-31 00:34:08.898487 +0000 UTC
Value: 1

When a call fails, it looks like:

Metric #1
Descriptor:
     -> Name: ftl.async_call.executed
     -> Description: the number of times that the controller tries executing an async call
     -> Unit: 1
     -> DataType: Sum
     -> IsMonotonic: true
     -> AggregationTemporality: Cumulative

NumberDataPoints #0
Data point attributes:
     -> ftl.async_call.execution.failure_mode: Str(async call failed)
     -> ftl.async_call.origin: Str(sub:echo.sub)
     -> ftl.async_call.time_since_scheduled_at_ms: Int(16)
     -> ftl.async_call.verb.ref: Str(echo.echoSinkOne)
     -> ftl.module.name: Str(echo)
     -> ftl.status.succeeded: Bool(false)
StartTimestamp: 2024-07-31 00:41:26.013892 +0000 UTC
Timestamp: 2024-07-31 00:41:41.014946 +0000 UTC
Value: 1

Issue: #2194

[deniseli]

Next metrics for async calls:

• Queue depth
• Latency as an actual metric, not just an attribute.