File

plugins/mod_limits.lua @ 11523:5f15ab7c6ae5

Statistics: Rewrite statistics backends to use OpenMetrics The metric subsystem of Prosody has had some shortcomings from the perspective of the current state-of-the-art in metric observability. The OpenMetrics standard [0] is a formalization of the data model (and serialization format) of the well-known and widely-used Prometheus [1] software stack. The previous stats subsystem of Prosody did not map well to that format (see e.g. [2] and [3]); the key reason is that it was trying to do too much math on its own ([2]) while lacking first-class support for "families" of metrics ([3]) and structured metric metadata (despite the `extra` argument to metrics, there was no standard way of representing common things like "tags" or "labels"). Even though OpenMetrics has grown from the Prometheus world of monitoring, it maps well to other popular monitoring stacks such as: - InfluxDB (labels can be mapped to tags and fields as necessary) - Carbon/Graphite (labels can be attached to the metric name with dot-separation) - StatsD (see graphite when assuming that graphite is used as backend, which is the default) The util.statsd module has been ported to use the OpenMetrics model as a proof of concept. An implementation which exposes the util.statistics backend data as Prometheus metrics is ready for publishing in prosody-modules (most likely as mod_openmetrics_prometheus to avoid breaking existing 0.11 deployments). At the same time, the previous measure()-based API had one major advantage: It is really simple and easy to use without requiring lots of knowledge about OpenMetrics or similar concepts. For that reason as well as compatibility with existing code, it is preserved and may even be extended in the future. However, code relying on the `stats-updated` event as well as `get_stats` from `statsmanager` will break because the data model has changed completely; in case of `stats-updated`, the code will simply not run (as the event was renamed in order to avoid conflicts); the `get_stats` function has been removed completely (so it will cause a traceback when it is attempted to be used). Note that the measure_*_event methods have been removed from the module API. I was unable to find any uses or documentation and thus deemed they should not be ported. Re-implementation is possible when necessary. [0]: https://openmetrics.io/ [1]: https://prometheus.io/ [2]: #959 [3]: #960
author Jonas Schäfer <jonas@wielicki.name>
date Sun, 18 Apr 2021 11:47:41 +0200
parent 10551:27b275633156
child 11560:3bbb1af92514
line wrap: on
line source

-- Because we deal with pre-authed sessions and streams we can't be host-specific
module:set_global();

local filters = require "util.filters";
local throttle = require "util.throttle";
local timer = require "util.timer";
local ceil = math.ceil;

local limits_cfg = module:get_option("limits", {});
local limits_resolution = module:get_option_number("limits_resolution", 1);

local default_bytes_per_second = 3000;
local default_burst = 2;

local rate_units = { b = 1, k = 3, m = 6, g = 9, t = 12 } -- Plan for the future.
local function parse_rate(rate, sess_type)
	local quantity, unit, exp;
	if rate then
		quantity, unit = rate:match("^(%d+) ?([^/]+)/s$");
		exp = quantity and rate_units[unit:sub(1,1):lower()];
	end
	if not exp then
		module:log("error", "Error parsing rate for %s: %q, using default rate (%d bytes/s)", sess_type, rate, default_bytes_per_second);
		return default_bytes_per_second;
	end
	return quantity*(10^exp);
end

local function parse_burst(burst, sess_type)
	if type(burst) == "string" then
		burst = burst:match("^(%d+) ?s$");
	end
	local n_burst = tonumber(burst);
	if not n_burst then
		module:log("error", "Unable to parse burst for %s: %q, using default burst interval (%ds)", sess_type, burst, default_burst);
	end
	return n_burst or default_burst;
end

-- Process config option into limits table:
-- limits = { c2s = { bytes_per_second = X, burst_seconds = Y } }
local limits = {};

for sess_type, sess_limits in pairs(limits_cfg) do
	limits[sess_type] = {
		bytes_per_second = parse_rate(sess_limits.rate, sess_type);
		burst_seconds = parse_burst(sess_limits.burst, sess_type);
	};
end

local default_filter_set = {};

function default_filter_set.bytes_in(bytes, session)
	local sess_throttle = session.throttle;
	if sess_throttle then
		local ok, _, outstanding = sess_throttle:poll(#bytes, true);
		if not ok then
			session.log("debug", "Session over rate limit (%d) with %d (by %d), pausing", sess_throttle.max, #bytes, outstanding);
			outstanding = ceil(outstanding);
			session.conn:pause(); -- Read no more data from the connection until there is no outstanding data
			local outstanding_data = bytes:sub(-outstanding);
			bytes = bytes:sub(1, #bytes-outstanding);
			timer.add_task(limits_resolution, function ()
				if not session.conn then return; end
				if sess_throttle:peek(#outstanding_data) then
					session.log("debug", "Resuming paused session");
					session.conn:resume();
				end
				-- Handle what we can of the outstanding data
				session.data(outstanding_data);
			end);
		end
	end
	return bytes;
end

local type_filters = {
	c2s = default_filter_set;
	s2sin = default_filter_set;
	s2sout = default_filter_set;
};

local function filter_hook(session)
	local session_type = session.type:match("^[^_]+");
	local filter_set, opts = type_filters[session_type], limits[session_type];
	if opts then
		if session.conn and session.conn.setlimit then
			session.conn:setlimit(opts.bytes_per_second);
			-- Currently no burst support
		else
			session.throttle = throttle.create(opts.bytes_per_second * opts.burst_seconds, opts.burst_seconds);
			filters.add_filter(session, "bytes/in", filter_set.bytes_in, 1000);
		end
	end
end

function module.load()
	filters.add_filter_hook(filter_hook);
end

function module.unload()
	filters.remove_filter_hook(filter_hook);
end

function module.add_host(module)
	local unlimited_jids = module:get_option_inherited_set("unlimited_jids", {});

	if not unlimited_jids:empty() then
		module:hook("authentication-success", function (event)
			local session = event.session;
			local session_type = session.type:match("^[^_]+");
			local jid = session.username .. "@" .. session.host;
			if unlimited_jids:contains(jid) then
				if session.conn and session.conn.setlimit then
					session.conn:setlimit(0);
					-- Currently no burst support
				else
					local filter_set = type_filters[session_type];
					filters.remove_filter(session, "bytes/in", filter_set.bytes_in);
					session.throttle = nil;
				end
			end
		end);
	end
end