Metric attributes with wrong value types fail badly

[alexmojaki]

import logfire

logfire.configure()

counter = logfire.metric_counter('my_counter')
counter.add(1, {'k': object()})

logs:

Failed to encode key k: Invalid type <class 'object'> of value <object object at 0x10864d320>
Traceback (most recent call last):
  File "/Users/alex/work/logfire/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/common/_internal/__init__.py", line 113, in _encode_attributes
    pb2_attributes.append(_encode_key_value(key, value))
                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/alex/work/logfire/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/common/_internal/__init__.py", line 94, in _encode_key_value
    return PB2KeyValue(key=key, value=_encode_value(value))
                                      ^^^^^^^^^^^^^^^^^^^^
  File "/Users/alex/work/logfire/.venv/lib/python3.12/site-packages/opentelemetry/exporter/otlp/proto/common/_internal/__init__.py", line 90, in _encode_value
    raise Exception(f"Invalid type {type(value)} of value {value}")
Exception: Invalid type <class 'object'> of value <object object at 0x10864d320>

This happens in the PeriodicExportingMetricReader thread and gives no way for the user to know the original offending code.

We should report this to OTEL if it isn't already and also decide what we want to do about it ourselves. I'm leaning towards warning the user and dropping the attribute.