diff --git a/app/vlselect/logsql/logsql.go b/app/vlselect/logsql/logsql.go
index 9cf15b7b7..f4b8a86e9 100644
--- a/app/vlselect/logsql/logsql.go
+++ b/app/vlselect/logsql/logsql.go
@@ -406,13 +406,20 @@ func ProcessLiveTailRequest(ctx context.Context, w http.ResponseWriter, r *http.
 	}
 	startOffset := startOffsetMsecs * 1e6
 
+	offsetMsecs, err := httputils.GetDuration(r, "offset", 1000)
+	if err != nil {
+		httpserver.Errorf(w, r, "%s", err)
+		return
+	}
+	offset := offsetMsecs * 1e6
+
 	ctxWithCancel, cancel := context.WithCancel(ctx)
 	tp := newTailProcessor(cancel)
 
 	ticker := time.NewTicker(refreshInterval)
 	defer ticker.Stop()
 
-	end := time.Now().UnixNano()
+	end := time.Now().UnixNano() - offset
 	start := end - startOffset
 	doneCh := ctxWithCancel.Done()
 	flusher, ok := w.(http.Flusher)
@@ -441,7 +448,7 @@ func ProcessLiveTailRequest(ctx context.Context, w http.ResponseWriter, r *http.
 			return
 		case <-ticker.C:
 			start = end - tailOffsetNsecs
-			end = time.Now().UnixNano()
+			end = time.Now().UnixNano() - offset
 		}
 	}
 }
diff --git a/docs/VictoriaLogs/CHANGELOG.md b/docs/VictoriaLogs/CHANGELOG.md
index 5320b7fbb..2c5074877 100644
--- a/docs/VictoriaLogs/CHANGELOG.md
+++ b/docs/VictoriaLogs/CHANGELOG.md
@@ -17,6 +17,7 @@ according to [these docs](https://docs.victoriametrics.com/victorialogs/quicksta
 
 * FEATURE: [`join` pipe](https://docs.victoriametrics.com/victorialogs/logsql/#join-pipe): add an ability to add prefix to all the log field names from the joined query, by using `| join by (<by_fields>) (<query>) prefix "some_prefix"` syntax.
 * FEATURE: [`_time` filter](https://docs.victoriametrics.com/victorialogs/logsql/#time-filter): allow specifying offset without time range. For example, `_time:offset 1d` matches all the logs until `now-1d` in the [`_time` field](https://docs.victoriametrics.com/victorialogs/keyconcepts/#time-field). This is useful when building graphs for time ranges with some offset in the past.
+* FEATURE: [`/select/logsql/tail` HTTP endpoint](): support for `offset` query arg, which can be used for delayed emission of matching logs during live tailing. Thanks to @Fusl for the initial idea and implementation in [this pull request](https://github.com/VictoriaMetrics/VictoriaMetrics/pull/7428).
 
 * BUGFIX: [HTTP querying APIs](https://docs.victoriametrics.com/victorialogs/querying/#http-api): properly take into account the `end` query arg when calculating time range for [`_time:duration` filter](https://docs.victoriametrics.com/victorialogs/logsql/#time-filter). Previously the `_time:duration` filter was treated as `_time:[now-duration, now)`, while it should be treated as `_time:[end-duration, end)`.
 
diff --git a/docs/VictoriaLogs/querying/README.md b/docs/VictoriaLogs/querying/README.md
index 743064c55..2a023c0a2 100644
--- a/docs/VictoriaLogs/querying/README.md
+++ b/docs/VictoriaLogs/querying/README.md
@@ -148,13 +148,29 @@ The `<query>` must conform the following rules:
   across multiple streams.
 
 Live tailing supports returning historical logs, which were ingested into VictoriaLogs before the start of live tailing. Pass `start_offset=<d>` query
-arg to `/select/logsql/tail` where `<d>` is the duration for returning historical logs. For example, the following request returns historical logs
+arg to `/select/logsql/tail` where `<d>` is the duration for returning historical logs. For example, the following command returns historical logs
 which were ingested into VictoriaLogs during the last hour, before starting live tailing:
 
 ```sh
 curl -N http://localhost:9428/select/logsql/tail -d 'query=*' -d 'start_offset=1h'
 ```
 
+Live tailing delays delivering new logs for one second, so they could be properly delivered from log collectors to VictoriaLogs.
+This delay can be changed via `offset` query arg. For example, the following command delays delivering new logs for 30 seconds:
+
+```sh
+curl -N http://localhost:9428/select/logsql/tail -d 'query=*' -d 'offset=30s'
+```
+
+Live tailing checks for new logs every second. The frequency for the check can be changed via `refresh_interval` query arg.
+For example, the following command instructs live tailing to check for new logs every 10 seconds:
+
+```sh
+curl -N http://localhost:9428/select/logsql/tail -d 'query=*' -d 'refresh_interval=10s'
+```
+
+It isn't recommended setting too low value for `refresh_interval` query arg, since this may increase load on VictoriaLogs without measurable benefits.
+
 **Performance tip**: live tailing works the best if it matches newly ingested logs at relatively slow rate (e.g. up to 1K matching logs per second),
 e.g. it is optimized for the case when real humans inspect the output of live tailing in the real time. If live tailing returns logs at too high rate,
 then it is recommended adding more specific [filters](https://docs.victoriametrics.com/victorialogs/logsql/#filters) to the `<query>`, so it matches less logs.