Tonight from 11:00 PM - 12:00 AM ET we need to make a few updates to our site.
We recommend finishing up your work before 11:00 PM and logging in after 12:00 AM.
Thank you for your patience while we make these important updates.
We have an application that synchs information between a local database and CTCT. This application has been running smoothly for years but in the past 2 months has started throwing socket errors on many of the calls to CTCT.
Last night we got a 504 gateway timeout.
Based on logs, we are seeing in cases where there is no response from CTCT after 20 seconds after sending a request. At which point we usually see the socket 10060 error. We are seeing these on scheduling campaigns and updating contacts.
20 seconds seems rather long.
Any ideas on resolving this?
Are you aware of these long response times?
Thanks for reaching out about this issue!
This is actually something that we are aware of and have been actively working on over the last several days. At present this issue is happening to a small fraction of a percentage of API calls related to contacts and campaigns and we've already made steps over the last few days to reduce the number of overall 504 errors. Because of the nature of these issues, you should see success even if you immediately retry the request on receiving a 504 error.
While I unfortunately cannot give an exact time-frame for when this will be fully resolved, I can say that it should be very soon.
If you have any questions, please feel free to ask!
I'm still getting 504 errors occasionally when POSTing campaigns. It's happened three or four times in the last couple of weeks. Can you provide an update on your progress eliminating these errors?
One issue that's more frustrating than the errors themselves - the requests are successful, as you noted, however, only parts are successful. When a campaign is created, the from and reply to email addresses do not save, and the campaign pulls in the account defaults.
Thank you for following up on this! Regarding the 504 timeouts, the answer depends on which version of the API you're working with. There is a known issue with the V1 API(Which is in a deprecated state currently) related to email campaigns that has been causing a very small percentage of campaign related API calls to fail (~0.03% was the most recent number). After spending a large amount of resources and coming up without any results, a decision was made at a higher level to focus our time and resources on continued development and support for our V2 API. I am definitely willing to try helping with individual cases, especially if there is any level of consistency in the timeouts.
If by chance you are seeing timeouts with the V2 API, then I would definitely like to get more details to determine the cause of that timeout. Additionally, this is the first I've heard of a case where the from and reply to email addresses are not correct on the resulting campaign. Is it possible to capture the payload of these requests that are timing out so that I can examine what is being sent vs what is created? This piece with the email addresses being incorrect warrants further investigation regardless of API version.
I have been using the v1 API. When v2 first came out I was having consistent issues with the HTML filters that were put into place, and I delayed my implementation of v2 until it was a bit more established. Fortunately the issues I was having with v2 seem to have been resolved over time. I've started implementing v2 and if the problem persists with the new API I'll be sure to let you know.
Thanks for your response.