Rate limiting change?

[Douglas Middleton]

Has there been a change to the rate limiting? As of a week ago we were able to do 5 parallel classifier deploy jobs and then do get status's every couple minutes on those with out any issue , and then when 1 of those finished start another job. And over the last week i noticed that any time we tried to do more then 1 deploy job at a time the post would fail fail with a timeout(even if i set the timeout to 20 minutes). Now today it seems like any time i do a deploy post i get a 429 even if i wait several hours between doing any api requests.

Thanks

[Stephen Chan]

I did just push changes to the API code including rate limiting, but I would've expected the changes to allow more requests if anything. I'll see if I can find out what's going on.

[Douglas Middleton]

Looks like it is working again for the most part. Though I am noticing a new error code on get status that i have never seen before 409.

{

    "errors": [

        {

            "detail": "This job isn't finished yet"

        }

    ]

}

This is not happening on all jobs. out of the 40 or so today it seems like just 1 is stuck in this state and has been that way for a while. Deploy job 81364

[Stephen Chan]

Darn, I introduced the exact same bug twice related to the status of failed job units, and I only fixed one of those yesterday. I'll fix the other one soon.

[Stephen Chan]

Should be fixed now.

As for the longer response times / timeouts in the past week, I don't know for sure what would have caused those. But as part of the update a couple of days ago, I worked on speeding up a couple parts of the API including submission of deploy jobs, so hopefully that helps. Besides that, I did a full restart of the server, since that sometimes solves mysterious performance issues (especially if it's been a while since the last restart, which it has).