diff --git a/EXTEND.md b/EXTEND.md
index f81d785..76863a5 100644
--- a/EXTEND.md
+++ b/EXTEND.md
@@ -42,7 +42,7 @@ type PoolPlugin struct {
 
 **`Inbox`** The shared channel that delivers received messages to the pool's consumer. All peers in the pool deliver to the same inbox channel. Workers must include their peer ID in each `InboxMessage`.
 
-**`Events`** The shared channel for lifecycle events. Workers emit `honeybee.EventConnected` and `honeybee.EventDisconnected` directly. All events include a timestamp in the `At` field.
+**`Events`** The shared channel for lifecycle events. Workers emit `honeybee.EventConnected`, `honeybee.EventDisconnected`, and `honeybee.EventDialFailed` directly. All events include a timestamp in the `At` field. `EventDialFailed` events also carry a non-nil `Err` field containing the dialer error.
 
 **`InboxCounter`** An atomic counter owned by the pool. Workers must increment this once for each message forwarded to `Inbox`. The pool reads it in `Stats()`.
 
@@ -66,7 +66,7 @@ The factory is set via `honeybee.WithWorkerFactory` on the pool config.
 
 ### Replacing the Worker
 
-Satisfy the `Worker` interface and register your implementation via `honeybee.WithWorkerFactory`. Your worker is responsible for the full connection lifecycle: dialing and redialing, managing connection state, forwarding received messages to `pool.Inbox`, emitting `EventConnected` and `EventDisconnected` to `pool.Events`, and incrementing `pool.InboxCounter` for each message forwarded.
+Satisfy the `Worker` interface and register your implementation via `honeybee.WithWorkerFactory`. Your worker is responsible for the full connection lifecycle: dialing and redialing, managing connection state, forwarding received messages to `pool.Inbox`, emitting `EventConnected`, `EventDisconnected`, and `EventDialFailed` to `pool.Events`, and incrementing `pool.InboxCounter` for each message forwarded.
 
 `DefaultWorker`'s source is the authoritative reference for how those responsibilities are met.
 
diff --git a/README.md b/README.md
index 7399be3..a971fc3 100644
--- a/README.md
+++ b/README.md
@@ -82,9 +82,11 @@ go func() {
 go func() {
     for ev := range pool.Events() {
         // ev.At             is the event timestamp
+        // ev.Err            is non-nil for EventDialFailed
         switch ev.Kind {
         case honeybee.EventConnected:
         case honeybee.EventDisconnected:
+        case honeybee.EventDialFailed:
         }
     }
 }()
@@ -104,7 +106,7 @@ After a disconnect, the worker waits for `ReconnectDelay` before attempting the
 
 `Send` returns `ErrConnectionUnavailable` during the gap between a disconnect and the next successful reconnect. Callers should wait for `EventConnected` before retrying and maintain their own write buffers if needed.
 
-Dial failures are handled internally by the worker's retry logic and documented in structured logs. These do not stop the pool; it continues retrying according to the connection's retry config.
+Each failed dial attempt emits `EventDialFailed` on `pool.Events()` with the dialer error in `ev.Err`. These do not stop the pool; it continues retrying according to the connection's retry config. `EventDialFailed` is only sent when the peer fails to connect, not when dialing is stopped internally.
 
 ## Server-Side Usage