Added non-agent-specific docs on span compression (#6438) (#6717)

Co-authored-by: Brandon Morelli <[email protected]>
Co-authored-by: Brandon Morelli <[email protected]>
Co-authored-by: Marc Lopez Rubio <[email protected]>
(cherry picked from commit 9ebd95e)

Co-authored-by: Benjamin Wohlwend <[email protected]>

docs/features.asciidoc

@@ -12,6 +12,7 @@
 * <<open-telemetry>>
 * <<log-correlation>>
 * <<cross-cluster-search>>
+* <<span-compression>>
 
 include::./apm-data-security.asciidoc[]
 
@@ -26,3 +27,5 @@ include::./open-telemetry.asciidoc[]
 include::./log-correlation.asciidoc[]
 
 include::./cross-cluster-search.asciidoc[]
+
+include::./span-compression.asciidoc[]

docs/span-compression.asciidoc

@@ -0,0 +1,65 @@
+[[span-compression]]
+=== Span compression
+
+In some cases, APM agents may collect large amounts of very similar or identical spans in a transaction.
+For example, this can happen if spans are captured inside of a loop, or in unoptimized SQL queries that use multiple queries instead of joins to fetch related data.
+In such cases, the upper limit of spans per transaction (by default, 500 spans) can be reached quickly, causing the agent to stop capturing potentially more relevant spans for a given transaction.
+
+Such repeated similar spans often aren't very relevant for themselves, especially if they are of very short duration.
+They also can clutter the UI, and cause processing and storage overhead.
+
+To address this problem, the APM agents can compress such spans into a single span.
+The compressed span retains most of the original span information, such as overall duration and the number of spans it represents.
+
+Regardless of the compression strategy, a span is elegible for compression if:
+
+- It has not propagated its trace context.
+- Is an _exit_ span (such as database query spans).
+- Its outcome is not `"failure"`.
+
+
+[float]
+[[span-compression-strategy]]
+=== Compression strategies
+
+The APM agent can select between two strategies to decide if two adjacent spans can be compressed.
+Both strategies have the benefit that only one previous span needs to be kept in memory.
+This is important to ensure that the agent doesn't require large amounts of memory to enable span compression.
+
+[float]
+[[span-compression-same]]
+==== Same-Kind strategy
+
+The agent selects this strategy if two adjacent spans have the same:
+
+ * span type
+ * span subtype
+ * `destination.service.resource` (e.g. database name)
+
+[float]
+[[span-compression-exact]]
+==== Exact-Match strategy
+
+The agent selects this strategy if two adjacent spans have the same:
+
+ * span name
+ * span type
+ * span subtype
+ * `destination.service.resource` (e.g. database name)
+
+[float]
+[[span-compression-settings]]
+=== Settings
+
+The agent has configuration settings to define upper thresholds in terms of span duration for both strategies.
+For the "Same-Kind" strategy, the limit is 5ms. For the "Exact-Match" strategy, the limit is 50ms.
+Spans with longer duration are not compressed. Please refer to the agent documentation for specifics.
+
+[float]
+[[span-compression-support]]
+=== Agent support
+
+Support for span compression is available in these agents:
+
+ * Go: https://www.elastic.co/guide/en/apm/agent/go/master/configuration.html#config-span-compression-exact-match-duration[`ELASTIC_APM_SPAN_COMPRESSION_EXACT_MATCH_MAX_DURATION`], https://www.elastic.co/guide/en/apm/agent/go/master/configuration.html#config-span-compression-same-kind-duration[`ELASTIC_APM_SPAN_COMPRESSION_SAME_KIND_MAX_DURATION`]
+ * Python: {apm-py-ref}/configuration.html#config-span-compression-exact-match-max_duration[`span_compression_exact_match_max_duration`], {apm-py-ref}/configuration.html#config-span-compression-same-kind-max-duration[`span_compression_same_kind_max_duration`]