diff --git a/waku/common/rate_limit/request_limiter.nim b/waku/common/rate_limit/request_limiter.nim
index 70fcc4905..7f33d0348 100644
--- a/waku/common/rate_limit/request_limiter.nim
+++ b/waku/common/rate_limit/request_limiter.nim
@@ -22,6 +22,8 @@ import
   libp2p/stream/connection,
   libp2p/utility
 
+import std/times except TimeInterval, Duration, seconds, minutes
+
 import ./[single_token_limiter, service_metrics, timed_map]
 
 export token_bucket, setting, service_metrics
@@ -76,8 +78,15 @@ template checkUsageLimit*(
     bodyWithinLimit, bodyRejected: untyped,
 ) =
   if t.checkUsage(proto, conn):
+    let requestStartTime = getTime().toUnixFloat()
     waku_service_requests.inc(labelValues = [proto, "served"])
+
     bodyWithinLimit
+
+    let requestDurationSec = getTime().toUnixFloat() - requestStartTime
+    waku_service_request_handling_duration_seconds.observe(
+      requestDurationSec, labelValues = [proto]
+    )
   else:
     waku_service_requests.inc(labelValues = [proto, "rejected"])
     bodyRejected
diff --git a/waku/common/rate_limit/service_metrics.nim b/waku/common/rate_limit/service_metrics.nim
index 339bf7f38..7d24d9530 100644
--- a/waku/common/rate_limit/service_metrics.nim
+++ b/waku/common/rate_limit/service_metrics.nim
@@ -17,3 +17,6 @@ proc setServiceLimitMetric*(service: string, limit: Option[RateLimitSetting]) =
     waku_service_requests_limit.set(
       limit.get().calculateLimitPerSecond(), labelValues = [service]
     )
+
+declarePublicHistogram waku_service_request_handling_duration_seconds,
+  "duration of non-relay service handling", ["service"]
diff --git a/waku/common/rate_limit/single_token_limiter.nim b/waku/common/rate_limit/single_token_limiter.nim
index 1b62114bf..da01f61bb 100644
--- a/waku/common/rate_limit/single_token_limiter.nim
+++ b/waku/common/rate_limit/single_token_limiter.nim
@@ -4,6 +4,8 @@
 
 import std/[options], chronos/timer, libp2p/stream/connection, libp2p/utility
 
+import std/times except TimeInterval, Duration
+
 import ./[token_bucket, setting, service_metrics]
 export token_bucket, setting, service_metrics
 
@@ -43,8 +45,15 @@ template checkUsageLimit*(
     bodyWithinLimit, bodyRejected: untyped,
 ) =
   if t.checkUsage(proto):
+    let requestStartTime = getTime().toUnixFloat()
     waku_service_requests.inc(labelValues = [proto, "served"])
+
     bodyWithinLimit
+
+    let requestDurationSec = getTime().toUnixFloat() - requestStartTime
+    waku_service_request_handling_duration_seconds.observe(
+      requestDurationSec, labelValues = [proto]
+    )
   else:
     waku_service_requests.inc(labelValues = [proto, "rejected"])
     bodyRejected