Librato reporter gets 400 errors from Librato if it has no data to send

[myronmarston]

In our system, the activity of our job queues is very bursty: every day around midnight, millions of jobs are enqueued and processed, and then there are times later in the day where there is no work to do. We've been getting a bunch of 400 errors from Librato with an error like this:

 {"errors":{"request":["Must contain a measurement"]}}

Looking at the Metriks librato reporter this makes sense: it submits all measurements that were taken in the last minute regardless of whether or not there were any. When there were none, it gets the 400 error from Librato.

It would be nice if the Metriks librato reporter was a bit smarter about this and would not try to submit a request with no metrics.

@eric, if you're on board with that idea I'll try to make time to submit a PR for this in the next week or so.

[eric]

Are you using the reporter from my experimental branch?

On Tue, Oct 1, 2013 at 1:44 PM, Myron Marston [email protected]
wrote:

In our system, the activity of our job queues is very bursty: every day around midnight, millions of jobs are enqueued and processed, and then there are times later in the day where there is no work to do. We've been getting a bunch of 400 errors from Librato with an error like this:

 {"errors":{"request":["Must contain a measurement"]}}

Looking at the Metriks librato reporter this makes sense: it submits all measurements that were taken in the last minute regardless of whether or not there were any. When there were none, it gets the 400 error from Librato.
It would be nice if the Metriks librato reporter was a bit smarter about this and would not try to submit a request with no metrics.

@eric, if you're on board with that idea I'll try to make time to submit a PR for this in the next week or so.

Reply to this email directly or view it on GitHub:
#42

[myronmarston]

Yep, we're using the enumerator-experimentation branch. We're currently on 0b6e337

[eric]

Cool. I have this fix for other versions of the reporter. Sorry for the oversight. Feel free to submit a pull requests. Otherwise I'll try to push a fix for it in the next day or so.

On Tue, Oct 1, 2013 at 1:51 PM, Myron Marston [email protected]
wrote:

Yep, we're using the enumerator-experimentation branch. We're currently on 0b6e337

Reply to this email directly or view it on GitHub:
#42 (comment)