Crash when reporter throws an error

[MaaKut]

I have a Hapi+Good configuration which reports to logstash via good-logstash-tcp. The problem is that whenever logstash instance becomes unavailable, an error is thrown by the reporter which in turn causes Good to crash the Hapi process.

Shouldn't Good handle errors like that? I can hardly imagine a situation where application crash in case of logging failure could be a desired behaviour.

[arb]

Yes it should. That's what these lines are for https://github.com/hapijs/good/blob/master/lib/monitor.js#L148-L151

If you can create a simple to reproduce case, I can investigate further.

[AdriVanHoudt]

@arb why not throw there?

[arb]

Because then it would crash the whole server.

[MaaKut]

If you can create a simple to reproduce case, I can investigate further.

Here's a code sample:

const good = require('good');

const server = new hapi.Server();

server.connection({
    port: 8081
});

server.route({
    method: 'GET',
    path: '/',
    handler: (rq, reply) => reply('ok')
});

server.register({
    register: good,
    options: {
        reporters: {
            consoleReporter: [
                { module: 'good-console' },
                'stdout'
            ],
            logstashReporter: [
                {
                    module: 'good-logstash-tcp',
                    args: [{
                        tlsOptions: {
                            host: 'localhost',
                            port: 5959
                        }
                    }]
                }
            ]
        }
    }
});

server.start(error => {
    if (error) {
        console.error('Server start failed:', error);
        throw error;
    }
    console.log(`Server is running: ${server.info.uri}...`);
});

The package.json is:

  "dependencies": {
    "good": "^7.3.0",
    "good-console": "^6.4.0",
    "good-logstash-tcp": "^0.2.1",
    "hapi": "^16.5.2"
  }
}

With this setup I get the following output:

events.js:163
      throw er; // Unhandled 'error' event
      ^

Error: Max retries reached, transport in silent mode, OFFLINE
    at Socket.<anonymous> (/.../logger-sample/node_modules/good-logstash-tcp/lib/logstash.js:178:36)
    at emitOne (events.js:96:13)
    at Socket.emit (events.js:191:7)
    at TCP._handle.close [as _onclose] (net.js:511:12)

...and the server stops

[AdriVanHoudt]

@arb (if you want we can discuss this elsewhere) is there a way to catch that? I'd prefer my server auto reboots so my logging works again then this weird state. What happens with the log stream, does it just voids?

[arb]

@MaaKut that's not really a simple server as it relies on modules outside the hapi organization. @AdriVanHoudt that makes sense, how would we handle that? Pass some kind of optional fail callback?

[MaaKut]

@arb The only external module here is good-logstash-tcp - it's just a reporter, whose crash brings down the whole server. Do you mean it throws an error in some proper way so that good is unable to handle it?

[AdriVanHoudt]

@arb either that or an option to tell it to throw, the callback would be cleaner though (so I can log the error still and then throw to crash and restart)

[lock[bot]]

This thread has been automatically locked due to inactivity. Please open a new issue for related bugs or questions following the new issue template instructions.