Load-balancers on IP Abuse database


As reported in https://wordpress.org/support/topic/api-wordpress-org-is-on-abuseip-db-blacklis/ it looks like some hosts have been confusing outgoing connections to api.wordpress.org as being incoming brute-force requests, and reporting it as such.

It appears that some hosts/networks may be using this as part of a network DROP rules causing WordPress sites being unable to connect to the APIAPI An API or Application Programming Interface is a software intermediary that allows programs to interact with each other and share data in limited, clearly defined ways.. I believe this might be the cause of some past issues where .251 was inaccessible but .252 was.

Can someone from systems request removal, specifically for https://www.abuseipdb.com/check/

#prio2 #lb

We’re going to be doing a HEAD request…

We’re going to be doing a HEAD request to https://api.wordpress.org/translations/core/1.0/?version=4.1 on the 4.1 about page, in JavaScript. While the heavy operations performed there are cached in memcache for 15 minutes, I think it may be prudent for us to (temporarily?) cache it at the load balancers to avoid being CPU-bound.

Suggestion: cache OPTIONS and HEAD requests to https://api.wordpress.org/translations/core/1.0/ for a period of time. A minute or 15 minutes is fine, but so is an hour.

#caching, #lb, #request