discourse/lib/middleware/request_tracker.rb

556 lines
17 KiB
Ruby
Raw Normal View History

# frozen_string_literal: true
require "method_profiler"
require "middleware/anonymous_cache"
require "http_user_agent_encoder"
class Middleware::RequestTracker
@@detailed_request_loggers = nil
@@ip_skipper = nil
# You can add exceptions to our app rate limiter in the app.yml ENV section.
# example:
#
# env:
# DISCOURSE_MAX_REQS_PER_IP_EXCEPTIONS: >-
# 14.15.16.32/27
# 216.148.1.2
#
STATIC_IP_SKIPPER =
ENV["DISCOURSE_MAX_REQS_PER_IP_EXCEPTIONS"]&.split&.map { |ip| IPAddr.new(ip) }
# register callbacks for detailed request loggers called on every request
# example:
#
# Middleware::RequestTracker.detailed_request_logger(->|env, data| do
# # do stuff with env and data
# end
def self.register_detailed_request_logger(callback)
MethodProfiler.ensure_discourse_instrumentation!
(@@detailed_request_loggers ||= []) << callback
end
def self.unregister_detailed_request_logger(callback)
@@detailed_request_loggers.delete(callback)
@detailed_request_loggers = nil if @@detailed_request_loggers.length == 0
end
# used for testing
def self.unregister_ip_skipper
@@ip_skipper = nil
end
# Register a custom `ip_skipper`, a function that will skip rate limiting
# for any IP that returns true.
#
# For example, if you never wanted to rate limit 1.2.3.4
#
# ```
# Middleware::RequestTracker.register_ip_skipper do |ip|
# ip == "1.2.3.4"
# end
# ```
def self.register_ip_skipper(&blk)
raise "IP skipper is already registered!" if @@ip_skipper
@@ip_skipper = blk
end
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
def self.ip_skipper
@@ip_skipper
end
2017-07-27 20:20:09 -05:00
def initialize(app, settings = {})
@app = app
end
def self.log_request(data)
if data[:is_api]
ApplicationRequest.increment!(:api)
elsif data[:is_user_api]
ApplicationRequest.increment!(:user_api)
elsif data[:track_view]
if data[:is_crawler]
ApplicationRequest.increment!(:page_view_crawler)
WebCrawlerRequest.increment!(data[:user_agent])
elsif data[:has_auth_cookie]
ApplicationRequest.increment!(:page_view_logged_in)
ApplicationRequest.increment!(:page_view_logged_in_mobile) if data[:is_mobile]
if data[:explicit_track_view]
# Must be a browser if it had this header from our ajax implementation
ApplicationRequest.increment!(:page_view_logged_in_browser)
ApplicationRequest.increment!(:page_view_logged_in_browser_mobile) if data[:is_mobile]
if data[:topic_id].present? && data[:current_user_id].present?
TopicsController.defer_topic_view(
data[:topic_id],
data[:request_remote_ip],
data[:current_user_id],
)
end
end
elsif !SiteSetting.login_required
ApplicationRequest.increment!(:page_view_anon)
ApplicationRequest.increment!(:page_view_anon_mobile) if data[:is_mobile]
if data[:explicit_track_view]
# Must be a browser if it had this header from our ajax implementation
ApplicationRequest.increment!(:page_view_anon_browser)
ApplicationRequest.increment!(:page_view_anon_browser_mobile) if data[:is_mobile]
if data[:topic_id].present?
TopicsController.defer_topic_view(data[:topic_id], data[:request_remote_ip])
end
end
end
end
# Message-bus requests may include this 'deferred track' header which we use to detect
# 'real browser' views.
if data[:deferred_track] && !data[:is_crawler]
if data[:has_auth_cookie]
ApplicationRequest.increment!(:page_view_logged_in_browser)
ApplicationRequest.increment!(:page_view_logged_in_browser_mobile) if data[:is_mobile]
if data[:topic_id].present? && data[:current_user_id].present?
TopicsController.defer_topic_view(
data[:topic_id],
data[:request_remote_ip],
data[:current_user_id],
)
end
elsif !SiteSetting.login_required
ApplicationRequest.increment!(:page_view_anon_browser)
ApplicationRequest.increment!(:page_view_anon_browser_mobile) if data[:is_mobile]
if data[:topic_id].present?
TopicsController.defer_topic_view(data[:topic_id], data[:request_remote_ip])
end
end
end
ApplicationRequest.increment!(:http_total)
status = data[:status]
if status >= 500
ApplicationRequest.increment!(:http_5xx)
elsif data[:is_background]
ApplicationRequest.increment!(:http_background)
elsif status >= 400
ApplicationRequest.increment!(:http_4xx)
elsif status >= 300
ApplicationRequest.increment!(:http_3xx)
elsif status >= 200
ApplicationRequest.increment!(:http_2xx)
end
end
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
def self.get_data(env, result, timing, request = nil)
2017-07-27 20:20:09 -05:00
status, headers = result
# result may be nil if the downstream app raised an exception
status = status.to_i
headers ||= {}
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
request ||= Rack::Request.new(env)
helper = Middleware::AnonymousCache::Helper.new(env, request)
# Since ActionDispatch::RemoteIp middleware is run before this middleware,
# we have access to the normalised remote IP based on ActionDispatch::RemoteIp::GetIp
#
# NOTE: Locally with MessageBus requests, the remote IP ends up as ::1 because
# of the X-Forwarded-For header set...somewhere, whereas all other requests
# end up as 127.0.0.1.
request_remote_ip = env["action_dispatch.remote_ip"].to_s
# Value of the discourse-track-view request header, set in `lib/ajax.js`
env_track_view = env["HTTP_DISCOURSE_TRACK_VIEW"]
# Was the discourse-track-view request header set to true? Likely
# set by our ajax library to indicate a page view.
explicit_track_view = status == 200 && %w[1 true].include?(env_track_view)
# An HTML response to a GET request is tracked implicitly
implicit_track_view =
status == 200 && !%w[0 false].include?(env_track_view) && request.get? && !request.xhr? &&
headers["Content-Type"] =~ %r{text/html}
# This header is sent on a follow-up request after a real browser loads up a page
# see `scripts/pageview.js` and `instance-initializers/page-tracking.js`
deferred_track_view = %w[1 true].include?(env["HTTP_DISCOURSE_DEFERRED_TRACK_VIEW"])
# This is treated separately from deferred tracking in #log_request, this is
# why deferred_track_view is not counted here.
track_view = !!(explicit_track_view || implicit_track_view)
# These are set in the same place as the respective track view headers in the client.
topic_id =
if deferred_track_view
env["HTTP_DISCOURSE_DEFERRED_TRACK_VIEW_TOPIC_ID"]
elsif explicit_track_view || implicit_track_view
env["HTTP_DISCOURSE_TRACK_VIEW_TOPIC_ID"]
end
auth_cookie = Auth::DefaultCurrentUserProvider.find_v0_auth_cookie(request)
auth_cookie ||= Auth::DefaultCurrentUserProvider.find_v1_auth_cookie(env)
has_auth_cookie = auth_cookie.present?
is_api ||= !!env[Auth::DefaultCurrentUserProvider::API_KEY_ENV]
is_user_api ||= !!env[Auth::DefaultCurrentUserProvider::USER_API_KEY_ENV]
is_message_bus = request.path.start_with?("#{Discourse.base_path}/message-bus/")
is_topic_timings = request.path.start_with?("#{Discourse.base_path}/topics/timings")
# Auth cookie can be used to find the ID for logged in users, but API calls must look up the
# current user based on env variables.
#
# We only care about this for topic views, other pageviews it's enough to know if the user is
# logged in or not, and we have separate pageview tracking for API views.
current_user_id =
if topic_id.present?
begin
(auth_cookie&.[](:user_id) || CurrentUser.lookup_from_env(env)&.id)
rescue Discourse::InvalidAccess => err
# This error is raised when the API key is invalid, no need to stop the show.
Discourse.warn_exception(
err,
message: "RequestTracker.get_data failed with an invalid API key error",
)
nil
end
end
h = {
status: status,
is_crawler: helper.is_crawler?,
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
has_auth_cookie: has_auth_cookie,
current_user_id: current_user_id,
topic_id: topic_id,
is_api: is_api,
is_user_api: is_user_api,
is_background: is_message_bus || is_topic_timings,
is_mobile: helper.is_mobile?,
track_view: track_view,
timing: timing,
queue_seconds: env["REQUEST_QUEUE_SECONDS"],
explicit_track_view: explicit_track_view,
deferred_track: deferred_track_view,
request_remote_ip: request_remote_ip,
}
if h[:is_background]
h[:background_type] = if is_message_bus
if request.query_string.include?("dlp=t")
"message-bus-dlp"
elsif env["HTTP_DONT_CHUNK"]
"message-bus-dontchunk"
else
"message-bus"
end
else
"topic-timings"
end
end
if h[:is_crawler]
user_agent = env["HTTP_USER_AGENT"]
user_agent = HttpUserAgentEncoder.ensure_utf8(user_agent) if user_agent
h[:user_agent] = user_agent
end
if cache = headers["X-Discourse-Cached"]
h[:cache] = cache
end
h
end
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
def log_request_info(env, result, info, request = nil)
# We've got to skip this on error ... its just logging
data =
begin
self.class.get_data(env, result, info, request)
rescue StandardError => err
Discourse.warn_exception(err, message: "RequestTracker.get_data failed")
# This is super hard to find if in testing, we should still raise in this case.
raise err if Rails.env.test?
nil
end
if data
2017-07-27 20:20:09 -05:00
if result && (headers = result[1])
2016-10-27 02:08:01 -05:00
headers["X-Discourse-TrackView"] = "1" if data[:track_view]
2016-10-27 00:50:56 -05:00
end
if @@detailed_request_loggers
@@detailed_request_loggers.each { |logger| logger.call(env, data) }
end
log_later(data)
end
end
def self.populate_request_queue_seconds!(env)
if !env["REQUEST_QUEUE_SECONDS"]
if queue_start = env["HTTP_X_REQUEST_START"]
queue_start =
if queue_start.start_with?("t=")
queue_start.split("t=")[1].to_f
else
queue_start.to_f / 1000.0
end
queue_time = (Time.now.to_f - queue_start)
env["REQUEST_QUEUE_SECONDS"] = queue_time
end
end
end
2017-11-27 23:47:20 -06:00
def call(env)
result = nil
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
info = nil
gc_stat_timing = nil
# doing this as early as possible so we have an
# accurate counter
::Middleware::RequestTracker.populate_request_queue_seconds!(env)
request = Rack::Request.new(env)
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
cookie = find_auth_cookie(env)
if error_details = rate_limit(request, cookie)
available_in, error_code, limit_on_id = error_details
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
message = <<~TEXT
Slow down, too many requests from this #{limit_on_id ? "user" : "IP address"}.
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
Please retry again in #{available_in} seconds.
Error code: #{error_code}.
TEXT
headers = {
"Content-Type" => "text/plain",
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
"Retry-After" => available_in.to_s,
"Discourse-Rate-Limit-Error-Code" => error_code,
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
}
if username = cookie&.[](:username)
headers["X-Discourse-Username"] = username
end
return 429, headers, [message]
end
if !cookie
if error_details = check_crawler_limits(env)
available_in, error_code = error_details
message = "Too many crawling requests. Error code: #{error_code}."
headers = {
"Content-Type" => "text/plain",
"Retry-After" => available_in.to_s,
"Discourse-Rate-Limit-Error-Code" => error_code,
}
return 429, headers, [message]
end
end
2017-11-27 23:47:20 -06:00
env["discourse.request_tracker"] = self
MethodProfiler.start
if SiteSetting.instrument_gc_stat_per_request
gc_stat_timing = GCStatInstrumenter.instrument { result = @app.call(env) }
else
result = @app.call(env)
end
info = MethodProfiler.stop
# possibly transferred?
2018-01-19 00:51:04 -06:00
if info && (headers = result[1])
headers["X-Runtime"] = "%0.6f" % info[:total_duration]
if GlobalSetting.enable_performance_http_headers
if redis = info[:redis]
headers["X-Redis-Calls"] = redis[:calls].to_s
headers["X-Redis-Time"] = "%0.6f" % redis[:duration]
end
if sql = info[:sql]
headers["X-Sql-Calls"] = sql[:calls].to_s
headers["X-Sql-Time"] = "%0.6f" % sql[:duration]
end
if queue = env["REQUEST_QUEUE_SECONDS"]
headers["X-Queue-Time"] = "%0.6f" % queue
end
end
end
if env[Auth::DefaultCurrentUserProvider::BAD_TOKEN] && (headers = result[1])
headers["Discourse-Logged-Out"] = "1"
end
2017-11-27 23:47:20 -06:00
result
ensure
if (limiters = env["DISCOURSE_RATE_LIMITERS"]) && env["DISCOURSE_IS_ASSET_PATH"]
limiters.each(&:rollback!)
env["DISCOURSE_ASSET_RATE_LIMITERS"].each do |limiter|
begin
limiter.performed!
rescue RateLimiter::LimitExceeded
# skip
end
end
end
if !env["discourse.request_tracker.skip"]
info.merge!(gc_stat_timing) if gc_stat_timing
log_request_info(env, result, info, request)
end
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
end
def log_later(data)
Scheduler::Defer.later("Track view") do
self.class.log_request(data) unless Discourse.pg_readonly_mode?
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
end
end
def find_auth_cookie(env)
min_allowed_timestamp = Time.now.to_i - (UserAuthToken::ROTATE_TIME_MINS + 1) * 60
cookie = Auth::DefaultCurrentUserProvider.find_v1_auth_cookie(env)
cookie if cookie && cookie[:issued_at] >= min_allowed_timestamp
2017-11-27 23:47:20 -06:00
end
def is_private_ip?(ip)
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
ip = IPAddr.new(ip)
!!(ip && (ip.private? || ip.loopback?))
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
rescue IPAddr::AddressFamilyError, IPAddr::InvalidAddressError
false
end
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
def rate_limit(request, cookie)
warn =
GlobalSetting.max_reqs_per_ip_mode == "warn" ||
GlobalSetting.max_reqs_per_ip_mode == "warn+block"
block =
GlobalSetting.max_reqs_per_ip_mode == "block" ||
GlobalSetting.max_reqs_per_ip_mode == "warn+block"
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
return if !block && !warn
ip = request.ip
if !GlobalSetting.max_reqs_rate_limit_on_private
return if is_private_ip?(ip)
end
return if @@ip_skipper&.call(ip)
return if STATIC_IP_SKIPPER&.any? { |entry| entry.include?(ip) }
ip_or_id = ip
limit_on_id = false
if cookie && cookie[:user_id] && cookie[:trust_level] &&
cookie[:trust_level] >= GlobalSetting.skip_per_ip_rate_limit_trust_level
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
ip_or_id = cookie[:user_id]
limit_on_id = true
end
limiter10 =
RateLimiter.new(
nil,
"global_limit_10_#{ip_or_id}",
GlobalSetting.max_reqs_per_ip_per_10_seconds,
10,
global: !limit_on_id,
aggressive: true,
error_code: limit_on_id ? "id_10_secs_limit" : "ip_10_secs_limit",
)
limiter60 =
RateLimiter.new(
nil,
"global_limit_60_#{ip_or_id}",
GlobalSetting.max_reqs_per_ip_per_minute,
60,
global: !limit_on_id,
error_code: limit_on_id ? "id_60_secs_limit" : "ip_60_secs_limit",
aggressive: true,
)
limiter_assets10 =
RateLimiter.new(
nil,
"global_limit_10_assets_#{ip_or_id}",
GlobalSetting.max_asset_reqs_per_ip_per_10_seconds,
10,
error_code: limit_on_id ? "id_assets_10_secs_limit" : "ip_assets_10_secs_limit",
global: !limit_on_id,
)
request.env["DISCOURSE_RATE_LIMITERS"] = [limiter10, limiter60]
request.env["DISCOURSE_ASSET_RATE_LIMITERS"] = [limiter_assets10]
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
if !limiter_assets10.can_perform?
if warn
limited_on = limit_on_id ? "user_id" : "ip"
Discourse.warn(
"Global asset rate limit exceeded for #{limited_on}: #{ip}: 10 second rate limit",
uri: request.env["REQUEST_URI"],
)
end
if block
return [
limiter_assets10.seconds_to_wait(Time.now.to_i),
limiter_assets10.error_code,
limit_on_id
]
end
end
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
begin
type = 10
limiter10.performed!
type = 60
limiter60.performed!
nil
rescue RateLimiter::LimitExceeded => e
if warn
limited_on = limit_on_id ? "user_id" : "ip"
Discourse.warn(
"Global rate limit exceeded for #{limited_on}: #{ip}: #{type} second rate limit",
uri: request.env["REQUEST_URI"],
)
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
end
if block
[e.available_in, e.error_code, limit_on_id]
FEATURE: Apply rate limits per user instead of IP for trusted users (#14706) Currently, Discourse rate limits all incoming requests by the IP address they originate from regardless of the user making the request. This can be frustrating if there are multiple users using Discourse simultaneously while sharing the same IP address (e.g. employees in an office). This commit implements a new feature to make Discourse apply rate limits by user id rather than IP address for users at or higher than the configured trust level (1 is the default). For example, let's say a Discourse instance is configured to allow 200 requests per minute per IP address, and we have 10 users at trust level 4 using Discourse simultaneously from the same IP address. Before this feature, the 10 users could only make a total of 200 requests per minute before they got rate limited. But with the new feature, each user is allowed to make 200 requests per minute because the rate limits are applied on user id rather than the IP address. The minimum trust level for applying user-id-based rate limits can be configured by the `skip_per_ip_rate_limit_trust_level` global setting. The default is 1, but it can be changed by either adding the `DISCOURSE_SKIP_PER_IP_RATE_LIMIT_TRUST_LEVEL` environment variable with the desired value to your `app.yml`, or changing the setting's value in the `discourse.conf` file. Requests made with API keys are still rate limited by IP address and the relevant global settings that control API keys rate limits. Before this commit, Discourse's auth cookie (`_t`) was simply a 32 characters string that Discourse used to lookup the current user from the database and the cookie contained no additional information about the user. However, we had to change the cookie content in this commit so we could identify the user from the cookie without making a database query before the rate limits logic and avoid introducing a bottleneck on busy sites. Besides the 32 characters auth token, the cookie now includes the user id, trust level and the cookie's generation date, and we encrypt/sign the cookie to prevent tampering. Internal ticket number: t54739.
2021-11-17 14:27:30 -06:00
else
nil
end
end
end
def check_crawler_limits(env)
slow_down_agents = SiteSetting.slow_down_crawler_user_agents
return if slow_down_agents.blank?
user_agent = HttpUserAgentEncoder.ensure_utf8(env["HTTP_USER_AGENT"])&.downcase
return if user_agent.blank?
return if !CrawlerDetection.crawler?(user_agent)
slow_down_agents
.downcase
.split("|")
.each do |crawler|
if user_agent.include?(crawler)
key = "#{crawler}_crawler_rate_limit"
limiter =
RateLimiter.new(nil, key, 1, SiteSetting.slow_down_crawler_rate, error_code: key)
limiter.performed!
break
end
end
nil
rescue RateLimiter::LimitExceeded => e
[e.available_in, e.error_code]
end
end