500 and 503 API errors


Yesterday we started seeing 500 and 503 response errors from Twitter with our DSP integration for many of our active campaigns. we’re seeing trouble with the statistics endpoints for campaign dashboard. This is for campaigns running from april 25 - may 16th

sample call: https://ads.twitter.com/accounts/<account_id>/campaigns_dashboard?campaign=<campaign_id>&endString=2016-05-16T11%3A20%3A22.000&segment=tweets&startString=2016-04-26T16%3A55%3A12.000

"Twitter API returned a 503 (Service Unavailable), Service unavailable due
to request timeout; please try the request again later"
and another one:
"Twitter API Error: Twitter API returned a 500 (Internal Server Error),
Internal Error”

I can provide specific campaign examples and links. just contact us via twitter-eng@amobee.com


More information on this ticket… we have been trying to set promoted tweet campaigns live. The endpoint to set the campaign live will throw an HTTP 500 or HTTP 503.


That’s a Twitter Ads UI URL, not an Ads API URL. Can you provide a sample API query? When did this start in UTC? What percentage of queries return a 5XX?


The original question was posted by a product manager. We are using the regular (non-batch) campaign update endpoint… documentation here ~ https://dev.twitter.com/ads/reference/put/accounts/%3Aaccount_id/campaigns/%3Acampaign_id

We saw this problem all day yesterday, starting at around 8AM Pacific Time. The last time someone intermittently tried to enable campaigns before this morning was at 5:30PM PT yesterday.

We then saw the same problem today for around half the work day (9AM PT - 1:30pm). I would say there’s about a 70% success rate based on the behavior I personally saw. I wasn’t attempting to hit the endpoint when the campaign manager’s contacted us about an inability to set the campaigns live at all, which was earlier today.

There is nothing unusual about the type of requests we’re making to the V1 ADS API campaign PUT endpoint. Here is sample data. Basically… we send up all the fields with the current value, and are toggling the “paused” field to mark the campaign ACTIVE.

{"account_id": "<account_id>",
 "campaign_id": "<the campaign id>",
"name": "<the campaign name>",
"start_time": "2016-05-16T04:00:00Z",
"end_time": "2016-08-01T03:59:00Z",
"standard_delivery": true,
"paused": false,
"total_budget_amount_local_micro": 10000000000,
"daily_budget_amount_local_micro": 77000000,
"frequency_cap": 5,
"duration_in_days": 30}

If you’d like the accountId, campaignName, campaignId, etc… please contact me directly via Elyssa Toda, and I can send you the client info via email. Thanks for your help!


Hi Chris,

After digging into this issue, it seems like these errors are only occurring for this specific advertiser. Could you confirm if this is true? I’ll attach more specific information in a whisper post (only viewable by you and other Twitter staff)


I know we’ve seen it for at least two separate advertisers in the past 24 hours. I’m definitely curious as to what the issue with those specific advertisers might be.

My guess is that they are very large accounts, possibly with a complex setup… or… it could be they have too many active campaigns already (over 200 across all their trafficking partners).


What email are notifications being sent to? I am not receiving updates even though I am ‘tracking’ the post.

Any update or guidance on what is causing these errors for the advertisers affected?


@amobeedev: to receive notifications by email for every post you need to be “watching” the topic. If you are just “tracking” it, you will receive notifications by email only if a person mentions you or replies to you directly (like I’m doing now). But if you created the topic, like this one, you should be “watching” the topic by default, and so you should receive email notifications for every new post. It sounds like you manually changed “watching” to “tracking”, though…