If we're asking AppSRE to provide a modern observability suite, we should be poised to make use of it. To that end, we should make sure that:

• Distributed traces are propagated correctly.
• Trace contexts/baggage are not leaked to external services.
• Trace spans are used throughout, ideally for both runtime and distributed traces.

Additionally, we should:

• Migrate direct usage of prometheus to otel/metric.
• Investigate using Sentry for trace exporting.
• Investigate enabling a flight recorder.