Solved: Task timed out after 30.00 seconds

Serverless Framework
1

Hello,
I have this function, but I keep getting time outs:

import { stripe } from '../../config';

export const getPlans = async ({ forceRefresh = false }) => {
  try {
    const existingPlans = await getPlansFromCache();
    if (!forceRefresh && shouldGetFromCache(existingPlans)) return JSON.parse(existingPlans.Body.toString('utf-8'));
    const result = await stripe.plans.list({ limit: 1000 });
    const plans = result.data.reduce((acc, plan) => {
      const [currency, , planCode] = plan.id.split('-');
      const { metadata = {} } = plan;
      if (!acc[planCode]) {
        return {
          ...acc,
          [planCode]: {
            planCode,
          },
        };
      }
      return {
        ...acc,
        [planCode]: {
          ...acc[planCode],
        },
      };
    }, []);
    return plans;
  } catch (err) {
    console.log('Err', err);
    return [];
  }
};
export default getPlans;

basically, I have a GraphQL endpoint which is executed from my Lambda function using POST method which then sends a request the stripe API to get all available plans, it writes it then writes it to S3.

here is part of my serverless.yml

  graphql:
    handler: build/main.graphql
    timeout: 30
    memory: 128
    events:
      - http:
          path: graphql
          method: post
          cors: true

any advice is much appreciated on how to fix this?

2

s3 was down yesterday. Perhaps this had something to do with it?

3

I don’t think so, as it is still not working and S3 is up.

const result = await stripe.plans.list({ limit: 1000 });

It groks at the above command - if I run it locally, all works as expected. The JSON payload is small!

4

Can you confirm you are able to reach the internet from the Lambda function?

This may sound silly, but if the Lambda is inside a VPC within a private subnet then you need a managed NAT instance (or roll your own) to access the internet - I fell into this trap just yesterday (took me ages to find out this little gem and fix it).

1 Like
5

hi yes, it is an issue with the lambda not being able to access the internet - a good resource i find is https://medium.com/@philippholly/aws-lambda-enable-outgoing-internet-access-within-vpc-8dd250e11e12#.hombvkfhy

1 Like
6

Awesome article - sorry I didn’t mention that I solved mine by using the internal dns of the EC2 that I wanted to access - prior to that I was using the public IP address - derp!

Hope you got your problem solved :smiley:

7

Hi, I have added my lambda function to my existing VPC and am able to return data back to my client, as detailed in the medium article.

Although I have an issue, which I am unclear as to why it is not working. On my VPC I have a mongo cluster, when I console.log the response from my lambda function, I am able to see it in the logs, but the curl client times out.

I am using,

curl -X POST -H "Authorization: Bearer _FQLs7c2k3E" -H "Content-Type: application/json" -H "Cache-Control: no-cache" -d '{ "query": "{ user { username } }" }' 'https://xxx.execute-api.us-east-1.amazonaws.com/dev/graphql'

other graphql queries work, when reading from S3 for example.

module.exports.graphql = (event, context, callback) => {
  const body = JSON.parse(event.body);
  const auth = authenticate(event.headers.Authorization);
  console.log('AUTH:', auth);
  if (!auth) { callback('ERROR: Authentication error'); return; }
  graphql(body.query, body.variables, auth)
    .then((response) => {
      console.log(response);
      callback(null, {
        statusCode: 200,
        body: JSON.stringify(response),
        headers: { 'Access-Control-Allow-Origin': '*' },
      });
    })
    .catch((err) => {
      console.log('ERROR: graphql', err);
      callback(err);
    });
};

here is the getUser, code.

export default ({ userId }) => new Promise(async (resolve, reject) => {
  try {
    console.log('try get user', userId);
    const user = await findUserById(userId);
    if (!user) return reject('User not found');
    return resolve({
      ...user,
      userId: String(user._id),
    });
  } catch (err) { return reject(err); }
});

here are the logs:

START RequestId: b187d7dd-0356-11e7-8adb-77c55d44b2a9 Version: $LATEST
2017-03-07 16:54:16.634 (+00:00)	b187d7dd-0356-11e7-8adb-77c55d44b2a9	AUTH: { role: 'user', userId: 'xxxx' }
2017-03-07 16:54:16.667 (+00:00)	b187d7dd-0356-11e7-8adb-77c55d44b2a9	try get user: xxxx
2017-03-07 16:54:17.335 (+00:00)	b187d7dd-0356-11e7-8adb-77c55d44b2a9	{ data: { user: { username: 'user@domain.tld' } } }

So my lambda function is connecting to the mongo cluster and getting the data from it, but it times out when returning it back to the client!

What am I missing here? Any advice is much appreciated.

8

OK, my issue was that the callback should return content.done(), this fixed it[quote=“khinester, post:7, topic:1422”]
callback(null, {
statusCode: 200,
body: JSON.stringify(response),
headers: { ‘Access-Control-Allow-Origin’: ‘*’ },
});
[/quote]

to

  content.done(null, {
    statusCode: 200,
    body: JSON.stringify(response),
    headers: { 'Access-Control-Allow-Origin': '*' },
  });

in https://github.com/serverless/serverless/issues/1036 is the old way!