From 6529aeba20e24c9a5f1f44c0d5f9eb32e59b08b6 Mon Sep 17 00:00:00 2001
From: overbit <2861984+overbit@users.noreply.github.com>
Date: Thu, 27 Oct 2022 21:48:39 +0100
Subject: [PATCH] doc: update documentation with new values

---
 docs/providers/aws/events/kafka.md | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/docs/providers/aws/events/kafka.md b/docs/providers/aws/events/kafka.md
index d06a95f6b44a..de2857528528 100644
--- a/docs/providers/aws/events/kafka.md
+++ b/docs/providers/aws/events/kafka.md
@@ -152,7 +152,8 @@ The Serverless Framework will automatically configure the most minimal set of IA
 
 You can set the `batchSize`, which effects how many messages can be processed in a single Lambda invocation. The default `batchSize` is 100, and the max `batchSize` is 10000.
 Likewise `maximumBatchingWindow` can be set to determine the amount of time the Lambda spends gathering records before invoking the function. The default is 0, but **if you set `batchSize` to more than 10, you must set `maximumBatchingWindow` to at least 1**. The maximum is 300.
-In addition, you can also configure `startingPosition`, which controls the position at which Lambda should start consuming messages from the topic. It supports two possible values, `TRIM_HORIZON` and `LATEST`, with `TRIM_HORIZON` being the default.
+In addition, you can also configure `startingPosition`, which controls the position at which Lambda should start consuming messages from the topic. It supports three possible values, `TRIM_HORIZON`, `LATEST` and `AT_TIMESTAMP`, with `TRIM_HORIZON` being the default.
+When `startingPosition` is configured as `AT_TIMESTAMP`, `startingPositionTimestamp` is also mandatory.
 
 In the following example, we specify that the `compute` function should have a `kafka` event configured with `batchSize` of 1000, `maximumBatchingWindow` of 30 seconds and `startingPosition` equal to `LATEST`.