Add sections for EDOT limitations (#400)

* Add EDOT limitations

* test

* Minor stylistic edits

* Apply comment

* Apply comments

* change 'Otel-native' phrasing

Author: alexandra5000

docs/reference/compatibility/features.md

@@ -22,8 +22,8 @@ The following table shows Elastic features and their level of support and compat
 | [Service Maps]                                              | [Compatible]     | [Supported]      |
 | [Distributed Tracing]                                       | [Compatible]     | [Supported]      |
 | [Head-based Sampling (HBS)]                                 | [Compatible]     | [Supported]      |
-| *[Tail-based Sampling (TBS)]*                               | [Compatible]     | [Not supported]  |
-| [Self-managed, OTel Collector-based TBS]                    | [Compatible]     | [Not supported]  |
+| [Tail-based Sampling (TBS)]                                 | [Compatible]^4^  | [Not supported]  |
+| [Self-managed, OTel Collector-based TBS]                    | [Compatible]^4^  | [Not supported]  |
 | TBS managed / hosted in {{ecloud}}                          | [Incompatible]   | [Not supported]  |
 | Runtime metrics                                             | [See language-specific overview](/reference/edot-sdks/index.md) | -                |
 | **Infrastructure Monitoring**                               | [Compatible]     | [Supported]      |
@@ -47,6 +47,8 @@ The following table shows Elastic features and their level of support and compat
 
 ^3^ Refer to [limitations on metrics ingestion](limitations.md#metrics-data-ingestion)
 
+^4^ While ingesting traces that have been sampled using TBS is technically possible, there are important limitations to consider. Refer to [Tail-based sampling (TBS) limitations](../compatibility/limitations.md#tail-based-sampling-tbs) for more information.
+
 [Incompatible]: nomenclature.md
 [Compatible]: nomenclature.md
 [Not supported]: nomenclature.md
@@ -57,4 +59,4 @@ The following table shows Elastic features and their level of support and compat
 [Head-based Sampling (HBS)]: docs-content://solutions/observability/apm/transaction-sampling.md#apm-head-based-sampling
 [Tail-based Sampling (TBS)]: docs-content://solutions/observability/apm/transaction-sampling.md#apm-tail-based-sampling
 [Self-managed, OTel Collector-based TBS]: https://opentelemetry.io/blog/2022/tail-sampling/
-[Time Series Data Streams]: docs-content://manage-data/data-store/data-streams/time-series-data-stream-tsds.md
+[Time Series Data Streams]: docs-content://manage-data/data-store/data-streams/time-series-data-stream-tsds.md

docs/reference/compatibility/limitations.md

@@ -16,7 +16,18 @@ products:
 
 The Elastic Distributions of OpenTelemetry (EDOT) come with a new way of ingesting data in OTel-native way and format. Elastic is continuously working on providing a great experience with OTel-native data within Elastic solutions, contributing popular Elastic features to the upstream OpenTelemetry projects and aligning concepts with OpenTelemetry.
 
-While EDOT and OTel-native data collection already covers most of the core Observability use cases, the following limitations apply compared to data collection with classic Elastic data collection mechanisms.
+While EDOT and OTel-native data collection already covers most of the core Observability use cases, the following limitations apply compared to data collection with classic Elastic data ingestion components.
+
+## When to use the classic Elastic Stack ingestion components instead of EDOT
+
+EDOT already supports most core observability use cases, but in some scenarios, you may prefer to use classic Elastic ingestion components, such as Elastic Agent, Elastic APM Agent or APM Server:
+
+* **Real user monitoring (RUM):** RUM ingestion and visualizations are not yet available for OTel-native data.
+* **Universal profiling:** This capability is currently only supported in the classic stack.
+* **Existing integrations and dashboards:** Many prebuilt Elastic integrations and dashboards are designed for ECS-formatted data and may not work as expected with the OpenTelemetry semantic conventions without customization.
+* **Ingest pipelines for structuring logs:** {{es}} ingest pipelines cannot directly parse OTel-native data with dotted field names without preprocessing. See [Centralized parsing and processing of data](#centralized-parsing-and-processing-of-data) for workarounds.
+* **Tail-based sampling (TBS):**  
+If you need the full tail-based sampling capabilities of APM Server, use APM Server with an Elasticsearch output. EDOT does not provide managed TBS. You can run TBS in a self-managed EDOT Collector or any upstream OTel Collector and ingest the sampled traces into Elastic with some caveats - refer to [Tail-based sampling limitations](#tail-based-sampling-tbs) for more information.
 
 Refer to [EDOT data streams compared to classic APM](../compatibility/data-streams.md) for an overview of how these ingestion paths differ.
 
@@ -71,6 +82,17 @@ Currently, there are limitations with visualizing language-specific runtime metr
 
 Runtime metrics can be ingested and used to create custom dashboards. As a temporary workaround users can create dashboards from the runtime metrics and attach them as custom dashboards to corresponding services.
 
+## Tail-based sampling (TBS)
+
+If you need the full tail-based sampling capabilities of APM Server, use APM Server with an Elasticsearch output. EDOT does not provide a managed TBS service.
+
+You can run tail-based sampling in a self-managed EDOT Collector or any upstream OTel Collector and ingest the sampled traces into Elastic, with these caveats:
+
+* **Metric accuracy:** Counts and rate metrics reflect sampled data, not total volumes. The Elastic APM backend cannot extrapolate totals because the `tailsamplingprocessor` does not send sampling probability metadata.
+* **Service map coverage:** Some edges between services may be missing.
+* **Impact on SLOs and alerts:** SLOs and alerts that depend on request volume can be biased by sampling.
+* **Operational complexity:** You are responsible for reliability, scaling, and tuning.
+
 ## Additional information
 
 For backwards compatibility reasons, Kubernetes metrics and host metrics are ingested twice, once in OTel format and once in Elastic Common Schema (ECS) format.