Updating ref docs to include the metrics bucket information for connection histograms

docs/core/diagnostics/built-in-metrics-aspnetcore.md

@@ -67,6 +67,8 @@ The time ends when:
 - All response data has been sent.
 - The context data structures for the request are being disposed.
 
+When using OpenTelemetry, the default buckets for this metric are set to [ 0.005, 0.01, 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1, 2.5, 5, 7.5, 10 ].
+
 <!-- Once we migrate this doc to https://github.com/dotnet/AspNetCore.Docs we can remove the following version info -->
 Available staring in: ASP.NET Core 8.0
 
@@ -240,6 +242,8 @@ Available staring in: ASP.NET Core 8.0
 | `server.port` | int | Server port number | `80`; `8080`; `443` | If the transport is `tcp` or `udp`. |
 | `tls.protocol.version` | string | TLS protocol version. | `1.2`; `1.3` | If the connection is secured with TLS. |
 
+As this metric is tracking the connection duration, and ideally http connections are used for multiple requests, the buckets should be longer than those used for request durations. For example, using [ 0.01, 0.02, 0.05, 0.1, 0.2, 0.5, 1, 2, 5, 10, 30, 60, 120, 300] provides an upper bucket of 5 mins.
+
 Available staring in: ASP.NET Core 8.0
 
 ### Instrument: `kestrel.rejected_connections`
@@ -323,6 +327,8 @@ Available staring in: ASP.NET Core 8.0
 | `server.port` | int | Server port number | `80`; `8080`; `443` | If the transport is `tcp` or `udp`. |
 | `tls.protocol.version` | string | TLS protocol version. | `1.2`; `1.3` | If the connection is secured with TLS. |
 
+When using OpenTelemetry, the default buckets for this metic are set to [ 0.005, 0.01, 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1, 2.5, 5, 7.5, 10 ].
+
 Available staring in: ASP.NET Core 8.0
 
 ### Instrument: `kestrel.active_tls_handshakes`
@@ -369,6 +375,8 @@ Available staring in: ASP.NET Core 8.0
 | `long_polling` | [Long Polling](/archive/msdn-magazine/2012/april/cutting-edge-long-polling-and-signalr) |
 | `web_sockets` | [WebSocket](https://datatracker.ietf.org/doc/html/rfc6455) |
 
+As this metric is tracking the connection duration, and ideally SignalR connections are durable, the buckets should be longer than those used for request durations. For example, using [0, 0.01, 0.02, 0.05, 0.1, 0.2, 0.5, 1, 2, 5, 10, 30, 60, 120, 300] provides an upper bucket of 5 mins.
+
 Available staring in: ASP.NET Core 8.0
 
 ### Instrument: `signalr.server.active_connections`

docs/core/diagnostics/built-in-metrics-system-net.md

@@ -46,6 +46,8 @@ Most errors when doing a DNS lookup throw a <xref:System.Net.Sockets.SocketExcep
 
 Socket exceptions with any other `SocketError` value are reported as `System.Net.Sockets.SocketException`.
 
+When using OpenTelemetry, the default buckets for this metric are set to [ 0.005, 0.01, 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1, 2.5, 5, 7.5, 10 ].
+
 Available starting in: .NET 8
 
 ## Meter: `System.Net.Http`
@@ -85,6 +87,8 @@ Available starting in: .NET 8
 
 This metric is only captured when <xref:System.Net.Http.HttpClient> is configured to use the default <xref:System.Net.Http.SocketsHttpHandler>.
 
+As this metric is tracking the connection duration, and ideally http connections are used for multiple requests, the buckets should be longer than those used for request durations. For example, using [ 0.01, 0.02, 0.05, 0.1, 0.2, 0.5, 1, 2, 5, 10, 30, 60, 120, 300 ] provides an upper bucket of 5 mins.
+
 Available starting in: .NET 8
 
 ### Instrument: `http.client.request.duration`
@@ -105,6 +109,8 @@ Available starting in: .NET 8
 
 HTTP client request duration measures the time the underlying client handler takes to complete the request. Completing the request includes the time up to reading response headers from the network stream. It doesn't include the time spent reading the response body.
 
+When using OpenTelemetry, the default buckets for this metric are set to [ 0.005, 0.01, 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1, 2.5, 5, 7.5, 10 ].
+
 Available starting in: .NET 8
 
 ### Instrument: `http.client.request.time_in_queue`