docs(observability): update example to signal-specific OTel env var names

[AryanKansagara]

Closes #352

What

Updates the example observability stack documentation to use the current OpenTelemetry environment variable names.

• examples/observability/otel-jaeger-prometheus/README.md: the variable table, the shell example, the .env example, and the closing sentence about when export is disabled now use TANGLE_OTEL_TRACE_EXPORTER_* and TANGLE_OTEL_METRIC_EXPORTER_*. Added one sentence noting that traces and metrics are configured independently. Corrected the documented TANGLE_ENV default from development to unknown.
• examples/observability/otel-jaeger-prometheus/docker-compose.yml: the header comment now lists the four current variables.

Why

#125 replaced the generic TANGLE_OTEL_EXPORTER_ENDPOINT and TANGLE_OTEL_EXPORTER_PROTOCOL with signal-specific variables and updated the code and tests, but the example directory was not updated. A user following the README gets no traces or metrics and no error, because configuration.resolve() returns None when neither of the current endpoint variables is set.

Verification

Loading the config with the variables from the updated .env block:

TANGLE_OTEL_TRACE_EXPORTER_ENDPOINT=http://localhost:4317 \
TANGLE_OTEL_METRIC_EXPORTER_ENDPOINT=http://localhost:4317 \
python -c "from cloud_pipelines_backend.instrumentation.opentelemetry._internal import configuration as c; r=c.resolve(); print(r.trace_exporter.endpoint, r.metrics.exporter.endpoint, r.service_name)"

prints http://localhost:4317 http://localhost:4317 tangle-unknown. The same command with the old variable names prints None.

git grep TANGLE_OTEL_EXPORTER_ now returns no results.

Documentation only, no code changes.