1 Answer
- Newest
- Most votes
- Most comments
0
If you configure a Lambda trigger from SQS, you can configure a batch size, and you can also configure a window size. The batch size indicates what is the maximum number of messages that your function will receive in each invocation and the window size indicates how long to wait until the batch gets full. If you do not specify a window size, your function will be invokes as soon as there will be messages in the queue.
Relevant content
- asked a year ago
- asked 24 days ago
- asked 10 months ago
- AWS OFFICIALUpdated a year ago
- AWS OFFICIALUpdated a year ago
- AWS OFFICIALUpdated a year ago
Thank you for your help!! it works! I set the window to 60 seconds and batch = 5 in my read sqs lambda function(Service B), since my lambda(Service A) is deployed in 5 different regions and when the message need to be sent from the SQS to the service B, now 4 records will be group together and send, and then the left 1 will be sent seprately , and sometimes 3 records will be grouped together and then the another 2 records are also group together then send from the SQS to the Service B, I wonder if it is impossible to guarantee that exactly 5 records, do not seprate any of those, can get from SQS and send to the Service B
Because of the distributed nature of SQS and the fact that you have very low number of messages, you can't guarantee this. If you will send a larger number of messages, most invocations will include the 5 records.
Did not try it, but if you switch to a FIFO queue, and use the same messgae group ID for all messages you send into the queue, it may also work.