count metric naming only applies to UpDownCounters

CHANGELOG.md

@@ -69,6 +69,8 @@ release.
      `net.sock.host.addr` to `server.socket.address` (since `net.sock.host.*` only applied to server instrumentation),
      `net.sock.host.port` to `server.socket.port` (similarly since `net.sock.host.*` only applied to server instrumentation),
      `http.client_ip` to `client.address`
+- Update `.count` metric naming convention so that it only applies to UpDownCounters
+  ([#3493](https://github.com/open-telemetry/opentelemetry-specification/pull/3493))
 
 ### Compatibility
 

specification/metrics/semantic_conventions/README.md

@@ -114,7 +114,7 @@ should not be pluralized, even if many data points are recorded.
 * `system.paging.faults`, `system.disk.operations`, and `system.network.packets`
 should be pluralized, even if only a single data point is recorded.
 
-#### Use `count` Instead of Pluralization
+#### Use `count` Instead of Pluralization for UpDownCounters
 
 If the value being recorded represents the count of concepts signified
 by the namespace then the metric should be named `count` (within its namespace).
@@ -125,6 +125,10 @@ to the processes then to represent the count of the processes we can have a metr
 `system.processes.count`. The suffix `count` here indicates that it is the count of
 `system.processes`.
 
+This rule should only be applied to UpDownCounters, since (monotonic) Counters have
+`_total` appended to their names when they are mapped to Prometheus, which would lead to
+`_count_total`.
+
 ## General Metric Semantic Conventions
 
 **Status**: [Mixed](../../document-status.md)