Trusted proxies, unauthenticated rate limits and various fixes

.env.template

@@ -324,6 +324,14 @@
 ## Set to the string "none" (without quotes), to disable any headers and just use the remote IP
 # IP_HEADER=X-Real-IP
 
+## Which addresses the header above is accepted from, defaults to "local".
+## Anyone able to reach Vaultwarden can set the header, and the client IP is used for the login and
+## admin rate limits, so it is only trusted when the request comes from a proxy listed here.
+## "local" accepts it from any non global address, which covers a reverse proxy running on the same
+## host or container network. Use "all" to accept it from anywhere, or list the addresses of your
+## proxy as IPs and CIDR ranges if it connects from a public address.
+# IP_HEADER_TRUSTED_PROXIES=local
+
 ## Icon service
 ## The predefined icon services are: internal, bitwarden, duckduckgo, google.
 ## To specify a custom icon service, set a URL template with exactly one instance of `{}`,
@@ -461,6 +469,13 @@
 ## Note that this applies to both the login and the 2FA, so it's recommended to allow a burst size of at least 2.
 # LOGIN_RATELIMIT_MAX_BURST=10
 
+## Number of seconds, on average, between requests from the same IP address to one of the rate limited
+## unauthenticated endpoints, like the password hint, the account recovery mails or accessing a Send.
+# UNAUTHENTICATED_RATELIMIT_SECONDS=60
+## Allow a burst of requests of up to this size, while maintaining the average indicated by `UNAUTHENTICATED_RATELIMIT_SECONDS`.
+## This budget is shared between all of those endpoints, so it is more lenient than the login one.
+# UNAUTHENTICATED_RATELIMIT_MAX_BURST=50
+
 ## BETA FEATURE: Groups
 ## Controls whether group support is enabled for organizations
 ## This setting applies to organizations.