diff --git a/packages/documentation/content/docs/router/configuration/environment-variables.mdx b/packages/documentation/content/docs/router/configuration/environment-variables.mdx
index 455ce170..ef36c7f8 100644
--- a/packages/documentation/content/docs/router/configuration/environment-variables.mdx
+++ b/packages/documentation/content/docs/router/configuration/environment-variables.mdx
@@ -2,7 +2,9 @@
 title: "Environment Variables"
 ---
 
-## Available Environment Variables
+To configure Hive Router using environment variables, you use the pre-defined environment variables listed below, or hook any configuration field to an environment variable.
+
+## Pre-defined Environment Variables
 
 The following environment variables are accepted by the router, and they override other
 configurations that can be set in the router's configuration file (`router.config.yaml`):
@@ -31,6 +33,34 @@ configurations that can be set in the router's configuration file (`router.confi
 > In cases where a configuration file (`router.config.yaml`) is used with environment variables at
 > the same time, environment variables will only override the specific configuration values.
 
+> In cases where the pre-defined environment variables are used, along with `from_env`, the pre-defined environment variable will be used
+> as they are considered preferred over the configuration file values.
+
+## Setting Any Config Value from an Environment Variable
+
+Beyond the fixed overrides above, any primitive field anywhere in the configuration file — strings,
+numbers, booleans — can be set from an environment variable using `from_env`, instead of a literal value:
+
+```yaml title="router.config.yaml"
+http:
+  port:
+    from_env: PORT
+```
+
+If the referenced environment variable is not set, the field falls back to its own default (or fails
+validation as usual if the field is required). You can also supply an inline fallback with `default`,
+which is used instead of the field's own default when the environment variable is unset:
+
+```yaml title="router.config.yaml"
+http:
+  port:
+    from_env: PORT
+    default: 4000
+```
+
+If you are using the plugin system, `from_env` also works inside a plugin's own `config` block, since it's resolved before the
+configuration is split up and handed to each plugin.
+
 ## Accessing Environment Variables in Expressions
 
 Beyond using environment variables to override configuration values at startup, you can also access
diff --git a/packages/documentation/content/docs/router/configuration/storages.mdx b/packages/documentation/content/docs/router/configuration/storages.mdx
index a405f16a..64d19af4 100644
--- a/packages/documentation/content/docs/router/configuration/storages.mdx
+++ b/packages/documentation/content/docs/router/configuration/storages.mdx
@@ -204,6 +204,15 @@ endpoint:
   are not dynamic in the way that header or routing expressions are.
 </Callout>
 
+For simple cases (no fallback logic beyond a single default), the more general
+[`from_env`](./environment-variables#setting-any-config-value-from-an-environment-variable) shorthand
+works here too, and on any other primitive field in the config:
+
+```yaml
+bucket:
+  from_env: S3_BUCKET
+```
+
 ## Provider examples
 
 ### Amazon S3 (production, EKS via IRSA)
diff --git a/packages/documentation/content/docs/router/customizations/plugin-system/hooks.mdx b/packages/documentation/content/docs/router/customizations/plugin-system/hooks.mdx
index 6d92ed2d..7e73fae8 100644
--- a/packages/documentation/content/docs/router/customizations/plugin-system/hooks.mdx
+++ b/packages/documentation/content/docs/router/customizations/plugin-system/hooks.mdx
@@ -952,6 +952,42 @@ async fn on_execute<'exec>(
 }
 ```
 
+**Report per-subgraph call durations**
+
+The request summary tracks `subgraph_calls_duration`, a map of subgraph name to the duration of
+every call made to it. Read it via
+[`get_current_summary()`](/docs/router/observability/logging#customizing-logs-from-plugins) in
+`on_end` and expose it however you like, e.g. as a response extension.
+
+```rust
+use hive_router::get_current_summary;
+use std::collections::BTreeMap;
+
+async fn on_execute<'exec>(
+    &'exec self,
+    payload: OnExecuteStartHookPayload<'exec>,
+) -> OnExecuteStartHookResult<'exec> {
+    payload.on_end(|mut end_payload: OnExecuteEndHookPayload<'exec>| {
+        let subgraph_calls: BTreeMap<String, u64> = get_current_summary()
+            .and_then(|summary| {
+                summary.subgraph_calls_duration.lock().ok().map(|durations| {
+                    durations
+                        .iter()
+                        .map(|(name, calls)| {
+                            let total_ms = calls.iter().map(|d| d.as_millis() as u64).sum();
+                            (name.clone(), total_ms)
+                        })
+                        .collect()
+                })
+            })
+            .unwrap_or_default();
+
+        end_payload.add_extension("subgraph_calls", subgraph_calls);
+        end_payload.proceed()
+    })
+}
+```
+
 #### Payload API Reference
 
 - [`OnExecuteStartHookPayload`](https://github.com/graphql-hive/router/blob/main/lib/executor/src/plugins/hooks/on_execute.rs#:~:text=struct%20OnExecuteStartHookPayload)
diff --git a/packages/documentation/content/docs/router/observability/logging.mdx b/packages/documentation/content/docs/router/observability/logging.mdx
index a1b0b3b4..88778aab 100644
--- a/packages/documentation/content/docs/router/observability/logging.mdx
+++ b/packages/documentation/content/docs/router/observability/logging.mdx
@@ -212,6 +212,7 @@ which makes filtering and routing precise. Common targets include:
 - `router::graphql_parsing` / `router::graphql_validation` / `router::graphql_normalization`
 - `router::coprocessor` / `router::plugin_system` — customizations
 - `router::jwt` / `router::authorization` — security
+- `router::config` — rotuer configuration loading
 
 ## Request correlation
 
@@ -256,8 +257,8 @@ Plugins can enrich the current request's logs through a few functions, callable
 - **`set_summary_message(message)`** — overrides the summary line's `message` field. The first call
   wins; later calls for the same request are no-ops.
 - **`get_current_summary()`** — reads the current request's in-progress summary (`operation_name`,
-  `status_code`, and so on), so a plugin can build a message from data only known once the request
-  is finishing.
+  `status_code`, `subgraph_calls_duration`, and so on), so a plugin can build a message from data
+  only known once the request is finishing.
 
 ```rust
 use hive_router::{get_current_summary, set_log_correlation, set_summary_message};