aws dynamodb client node batchwrite limit

The worst case scenario here is that a failed DynamoDB call takes over 50 sec before responding to your Lambda function. Starting with 50 ms for the first retry, then 100 ms, 200 ms, 400 ms all the way to 25600 ms for the last retry. However for calls to DynamoDB, the client auto-retries up to 10 times. It waits 100 ms before the first retry, 200 ms before the second, and 400 ms before the third attempt. For calls to most AWS services, the client auto-retries up to 3 times. It uses an exponential backoff strategy while retrying. By default, the AWS Node.js SDK client auto-retries such errors. However, the response usually specifies if the error can be retried. Import AWS from ' aws-sdk ' const dynamoDb = new AWS. In our Lambda function, change the timeout by using the following: So the first thing we need to do is to change the connection timeout for our DynamoDB client. As we talked about the issue above, our Lambda functions were timing out because the DynamoDB calls were taking too long. However, that is not what happens by default. We want the DynamoDB calls to timeout before Lambda and API Gateway timeout. If DynamoDB times out, we want it to fail quickly and retry again.

Hence, Lambda should have a timeout of lesser than 29 sec.Īnd it only makes sense that DynamoDB’s connection timeout is set to be lesser than that of Lambda.

API Gateway has a maximum timeout setting of 29 sec.

On the other hand, for Lambda and API Gateway: Sets the socket to timeout after a set amount of inactivity on the socket. This timeout has no effect once a socket connection has been established. This sets the socket to timeout after failing to establish a connection with the server after a set amount of milliseconds.

The first reason DynamoDB calls are taking long is because the HTTP requests that the AWS SDK is making to DynamoDB is timing out. There are two common ways DynamoDB calls might timeout. Thus, the next step was to figure out how to handle these timeouts so that our users weren’t being adversely affected by them. We confirmed with Mike Apted that this is a farily normal timeout rate for DynamoDB calls. Roughly 0.00002% of all our DynamoDB calls were timing out. We observed timeouts across BatchGet, Get, Query, and Update calls. However there were no specific types of calls that were timing out. We used AWS X-Ray to trace the calls and found that the DynamoDB call seemed to be taking abnormally long. We noticed that a certain percentage of our Node.js Lambda function calls were timing out. Let’s start by looking at the issue itself. Hopefully sharing our experience ends up being helpful to other folks. But we recently had to work through this issue. This is a fairly specific issue related to Node.js Lambda functions calling DynamoDB. In this post we are going to look at how to debug DynamoDB timeouts in Serverless apps.

Aws dynamodb client node batchwrite limit