Last week I had a problem with the purge API. During a deploy, we purge several URLs. This had been working for years, but, last week purge requests were getting stuck for a minute or so and then returned successfully. Specifically, some purges (10 or so) returned fast but then, one of them took a lot to finish. This caused our deploy to take several hours.
This only happened on our servers though, not on our local machines. I suspect it has to do with our servers having faster network speed. Maybe there is a rate limit?
Also, we managed to fix it using a connection pool to make the requests. That makes me think this is also related to a max number of connections limit on fastly side. I am not sure though.
Maybe you can clarify this to me please. I have looked at the docs; rate and connection limits are mentioned, something around 1000 requests per hour and 200 connections, but we shouldn't be hitting that limit.
Please sign in to leave a comment.