In "GlobalSetting.redirect_avatar_requests" mode, when the application gets
an avatar request it returns a "redirect" to the S3 CDN.
This shields the application from caching avatars and downloading from S3.
However clients will make 2 requests per avatar. (one to get redirect,
second to get avatar)
A one hour cache on a redirect means there may be an increase in CDN
traffic, given more clients will ask for the redirect every hour.
This may also lead to an increase in origin requests to the application.
To mitigate lets cache the CDN URL for 1 day.
The downside is that any changes to S3 CDN need extra care to allow for
the extra 1 day delay. (leave data around for 1 extra day)
When uploads are stored on S3, by default Discourse will fetch the avatars and proxy them through to the requesting client. This is simple, but it can lead to significant inbound/outbound network load in the hosting environment.
This commit adds an optional redirect_avatar_requests GlobalSetting. When enabled, requests for user avatars will be redirected to the S3 asset instead of being proxied. This adds an extra round-trip for clients, but it should significantly reduce server load. To mitigate that extra round-trip for clients, a CDN with 'follow redirect' capability could be used.
It's very easy to forget to add `require 'rails_helper'` at the top of every core/plugin spec file, and omissions can cause some very confusing/sporadic errors.
By setting this flag in `.rspec`, we can remove the need for `require 'rails_helper'` entirely.
Extracted commonly used spec helpers into spec/support/uploads_helpers.rb, removed unused stubs and let definitions. Makes it easier to write new S3-related specs without copy and pasting setup steps from other specs.
This change both speeds up specs (less strings to allocate) and helps catch
cases where methods in Discourse are mutating inputs.
Overall we will be migrating everything to use #frozen_string_literal: true
it will take a while, but this is the first and safest move in this direction
Co-authored-by: Sam Saffron <sam.saffron@gmail.com>
Co-authored-by: David Taylor <david@taylorhq.com>
This gives more control over the request. In particular we can easily
lookup DNS dynamically, instead of only upon NGINX startup.
Previously, NGINX was looking up IP for the letter avatar service and
caching the CDN IP address, this caused issues if CDN changed IP, in
which letter avatars would be broken till a container restarted.
NGINX config has been updated to add caching. This change will require
a container rebuild.
The proxy will now function in development environments, so the patch
for `letter_avatar_proxy` has been removed.
Previously we killed caching on old avatars cause we kept serving blank
this meant we would front many more avatar requests after a version change
This change ensures all old avatars do not cause a flood of requests on the
server
In some cases Akami was holding tight to these invalid avatars,
to avoid this happening we explain the avatar image is ancient
then when a new upload is added it automatically is older than
this.
