Seeing lots of 415 with a few versions of this agent

[jdotpz]

<182>Sep 13 17:26:26 i-d35187be app=makevalet-production: {"name":"newrelic","hostname":"i-d35187be","pid":28825,"component":"error_tracer","level":40,"err":{"message":"Got HTTP 415 in response to error_data.","name":"Error","stack":"Error: Got HTTP 415 in response to error_data.\n at DataSender.onCollectorResponse (/var/www/makevalet/node_modules/newrelic/lib/collector/data-sender.js:148:22)\n at ClientRequest.EventEmitter.emit (events.js:95:17)\n at HTTPParser.parserOnIncomingClient as onIncoming\n at HTTPParser.parserOnHeadersComplete as onHeadersComplete\n at Socket.socketOnData as ondata\n at TCP.onread (net.js:510:27)"},"msg":"Adding 2 errors for next harvest because harvest failed. Reason:","time":"2013-09-13T17:26:26.263Z","v":0}

f

[othiym23]

Thanks for the report! Is there a chance that there's a proxy somewhere between your application and New Relic's servers? The only time we send back a 415 is if there's an error deserializing the JSON on our end, and since the JSON is coming out of Node, it's highly unlikely to be serialized incorrectly.

If you want to help us figure out what's going on, open a New Relic support ticket, mentioning this issue and me, and we'll see if we can figure out what's going on.

[othiym23]

Have you seen a recurrence of this since the release of v0.11.0? We made a number of tweaks related to handling 413 and 415 status codes coming from New Relic, and I'm curious if you're still seeing this behavior.

[othiym23]

Closing because handling of status code 415 returned by New Relic has changed substantially since this issue was opened.

[aladagemre]

I'm having this problem with 1.3.2 version. New Relic web panel does not get the data from Heroku.

UPDATE: With the help of othiym23, we found that it was due to UTF8 char in the app name. So prefer to use ASCII chars in your app name.

2014-03-13T20:20:51.344647+00:00 app[web.1]: {"name":"newrelic","hostname":"...","pid":16,"component":"new_relic_call","level":20,"msg":"Finished receiving data back from the collector for connect.","time":"2014-03-13T20:20:51.344Z","v":0}
2014-03-13T20:20:51.344718+00:00 app[web.1]: {"name":"newrelic","hostname":"...","pid":16,"component":"new_relic_response","level":20,"msg":"Got 415 as a response code from the collector.","time":"2014-03-13T20:20:51.344Z","v":0}
2014-03-13T20:20:51.345113+00:00 app[web.1]: {"name":"newrelic","hostname":"...","pid":16,"component":"collector_api","level":20,"msg":"Failed attempting to connect to New Relic, waiting 120s to retry.","time":"2014-03-13T20:20:51.345Z","v":0}
2014-03-13T20:20:51.516788+00:00 app[web.1]: {"name":"newrelic","hostname":"...","pid":10,"component":"new_relic_response","level":20,"msg":"Got 415 as a response code from the collector.","time":"2014-03-13T20:20:51.516Z","v":0}
2014-03-13T20:20:51.516720+00:00 app[web.1]: {"name":"newrelic","hostname":"...","pid":10,"component":"new_relic_call","level":20,"msg":"Finished receiving data back from the collector for connect.","time":"2014-03-13T20:20:51.516Z","v":0}
2014-03-13T20:20:51.517264+00:00 app[web.1]: {"name":"newrelic","hostname":"...","pid":10,"component":"collector_api","level":20,"msg":"Failed attempting to connect to New Relic, waiting 120s to retry.","time":"2014-03-13T20:20:51.517Z","v":0}

[othiym23]

Can you try raising the logging level to trace on one host for a little while so we can see what data is getting sent to New Relic's servers? If you're getting 415 responses on the connection handshake (which it looks like is what's happening), then the initial connect message sent by New Relic for Node has something in it that New Relic's back end can't deserialize, and that's not something I'd expect, as Node sends valid JSON by default.

If you do gather those logs, can you send them to [email protected]? That's the easiest way to get that information to me.