Stream ended unexpectedly with status null and empty body about node-apn HOT 8 CLOSED

Is it happening on your production server or your local machine? Any chance your internet connection is hanging?

from node-apn.

Happen on production servers on AWS. So, I think that internet is good there

from node-apn.

const provider = new apn.Provider(config.apns);

Can you provide more details about what that config is? As much detail, e.g. are proxies being used, what keys and what values (where not credentials), etc.

What node.js version are you using in AWS?

How many notifications per second? What is the typical cpu usage for node in that application (e.g. if it's close to 100% or there's a high cpu time per tick requests might time out)

How often does this happen per hour in a single process?
Are there also error messages about reconnecting?

Is this a long-lived application or a short-lived script (not sure if it'd matter, but lambda seems to have had issues in other web servers for reasons I'm not familiar with)? aws/aws-sdk-js#3591 mentions a similar error message in a completely different context

from node-apn.

Ok, I'll provide you that information below.

Can you provide more details about what that config is? As much detail, e.g. are proxies being used, what keys and what values (where not credentials), etc.

My config.apns content. I'm not using proxy.

// p8 certificate
const key = `
-----BEGIN PRIVATE KEY-----
xxxxxxxx
-----END PRIVATE KEY-----
`;

export default {
  token: {
    key,
    keyId: 'xxxxxx',
    teamId: 'xxxxxx',
  },
  production: true,
  requestTimeout: 15000,
};

What node.js version are you using in AWS?

nodejs: v14.17.3

How many notifications per second? What is the typical cpu usage for node in that application (e.g. if it's close to 100% or there's a high cpu time per tick requests might time out)

About ~40/second notifications.
CPU usage: I don't know since I'm running the script on Lambda and I haven't that information.

How often does this happen per hour in a single process?

Sometimes could pass a couple of days without any issue. And sometimes, like today (Aug 9th, 2021), about 5 times along the day.

Are there also error messages about reconnecting?

Mmmm no, I didn't see any error related to reconnecting. But sometimes, the error is a timeout error instead the mentioned in the issue (apn write timeout)

Is this a long-lived application or a short-lived script (not sure if it'd matter, but lambda seems to have had issues in other web servers for reasons I'm not familiar with)? aws/aws-sdk-js#3591 mentions a similar error message in a completely different context

short-time and I running the script on Lambda, I'll check that issue

from node-apn.

Are any function to reconnect the client/provider when I detect that error? Or this happen be automatically?

from node-apn.

I can reproduce this issue using latest version 5.1.0 in this environment:

• Local macos: always working for node v14 and v16
• AWS node v16: working
• AWS node v14: Stream ended unexpectedly with status null and empty body

The AWS I was deploying in a docker container behind LB.

from node-apn.

I'm using node v14 since node v16 is not supported by lambda yet

from node-apn.

I'm closing this as it seems to be resolved. Feel free to comment if you have any questions and we can re-open this issue.

from node-apn.