@@ -28,6 +28,11 @@ class FirehoseClient extends AbstractApi
2828 * delivery stream. For more information about limits and how to request an increase, see Amazon Kinesis Data Firehose
2929 * Limits [^1].
3030 *
31+ * Kinesis Data Firehose accumulates and publishes a particular metric for a customer account in one minute intervals.
32+ * It is possible that the bursts of incoming bytes/records ingested to a delivery stream last only for a few seconds.
33+ * Due to this, the actual spikes in the traffic might not be fully visible in the customer's 1 minute CloudWatch
34+ * metrics.
35+ *
3136 * You must specify the name of the delivery stream and the data record when using PutRecord. The data record consists
3237 * of a data blob that can be up to 1,000 KiB in size, and any kind of data. For example, it can be a segment from a log
3338 * file, geographic location data, website clickstream data, and so on.
@@ -40,8 +45,12 @@ class FirehoseClient extends AbstractApi
4045 * The `PutRecord` operation returns a `RecordId`, which is a unique string assigned to each record. Producer
4146 * applications can use this ID for purposes such as auditability and investigation.
4247 *
43- * If the `PutRecord` operation throws a `ServiceUnavailableException`, back off and retry. If the exception persists,
44- * it is possible that the throughput limits have been exceeded for the delivery stream.
48+ * If the `PutRecord` operation throws a `ServiceUnavailableException`, the API is automatically reinvoked (retried) 3
49+ * times. If the exception persists, it is possible that the throughput limits have been exceeded for the delivery
50+ * stream.
51+ *
52+ * Re-invoking the Put API operations (for example, PutRecord and PutRecordBatch) can result in data duplicates. For
53+ * larger data assets, allow for a longer time out before retrying Put API operations.
4554 *
4655 * Data records sent to Kinesis Data Firehose are stored for 24 hours from the time they are added to a delivery stream
4756 * as it tries to send the records to the destination. If the destination is unreachable for more than 24 hours, the
@@ -84,6 +93,11 @@ public function putRecord($input): PutRecordOutput
8493 * producer than when writing single records. To write single data records into a delivery stream, use PutRecord.
8594 * Applications using these operations are referred to as producers.
8695 *
96+ * Kinesis Data Firehose accumulates and publishes a particular metric for a customer account in one minute intervals.
97+ * It is possible that the bursts of incoming bytes/records ingested to a delivery stream last only for a few seconds.
98+ * Due to this, the actual spikes in the traffic might not be fully visible in the customer's 1 minute CloudWatch
99+ * metrics.
100+ *
87101 * For information about service quota, see Amazon Kinesis Data Firehose Quota [^1].
88102 *
89103 * Each PutRecordBatch request supports up to 500 records. Each record in the request can be as large as 1,000 KB
@@ -117,8 +131,11 @@ public function putRecord($input): PutRecordOutput
117131 * processing. This minimizes the possible duplicate records and also reduces the total bytes sent (and corresponding
118132 * charges). We recommend that you handle any duplicates at the destination.
119133 *
120- * If PutRecordBatch throws `ServiceUnavailableException`, back off and retry. If the exception persists, it is possible
121- * that the throughput limits have been exceeded for the delivery stream.
134+ * If PutRecordBatch throws `ServiceUnavailableException`, the API is automatically reinvoked (retried) 3 times. If the
135+ * exception persists, it is possible that the throughput limits have been exceeded for the delivery stream.
136+ *
137+ * Re-invoking the Put API operations (for example, PutRecord and PutRecordBatch) can result in data duplicates. For
138+ * larger data assets, allow for a longer time out before retrying Put API operations.
122139 *
123140 * Data records sent to Kinesis Data Firehose are stored for 24 hours from the time they are added to a delivery stream
124141 * as it attempts to send the records to the destination. If the destination is unreachable for more than 24 hours, the
0 commit comments