File

util/timer.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 11264:2cdcf55c6dd5
child 12975:d10957394a3c
line wrap: on
line source

-- Prosody IM
-- Copyright (C) 2008-2010 Matthew Wild
-- Copyright (C) 2008-2010 Waqas Hussain
--
-- This project is MIT/X11 licensed. Please see the
-- COPYING file in the source package for more information.
--

local indexedbheap = require "util.indexedbheap";
local log = require "util.logger".init("timer");
local server = require "net.server";
local get_time = require "util.time".now
local type = type;
local debug_traceback = debug.traceback;
local tostring = tostring;
local xpcall = require "util.xpcall".xpcall;
local math_max = math.max;
local pairs = pairs;

if server.timer then
	-- The selected net.server implements this API, so defer to that
	return server.timer;
end

local _ENV = nil;
-- luacheck: std none

local _add_task = server.add_task;

local _server_timer;
local _active_timers = 0;
local h = indexedbheap.create();
local params = {};
local next_time = nil;
local function _traceback_handler(err) log("error", "Traceback[timer]: %s", debug_traceback(tostring(err), 2)); end
local function _on_timer(now)
	local peek;
	local readd;
	while true do
		peek = h:peek();
		if peek == nil or peek > now then break; end
		local _, callback, id = h:pop();
		local param = params[id];
		params[id] = nil;
		--item(now, id, _param);
		local success, err = xpcall(callback, _traceback_handler, now, id, param);
		if success and type(err) == "number" then
			if readd then
				readd[id] = { callback, err + now };
			else
				readd = { [id] = { callback, err + now } };
			end
			params[id] = param;
		end
	end

	if readd then
		for id,timer in pairs(readd) do
			h:insert(timer[1], timer[2], id);
		end
		peek = h:peek();
	end

	if peek ~= nil and _active_timers > 1 and peek == next_time then
		-- Another instance of _on_timer already set next_time to the same value,
		-- so it should be safe to not renew this timer event
		peek = nil;
	else
		next_time = peek;
	end

	if peek then
		-- peek is the time of the next event
		return peek - now;
	end
	_active_timers = _active_timers - 1;
end
local function add_task(delay, callback, param)
	local current_time = get_time();
	local event_time = current_time + delay;

	local id = h:insert(callback, event_time);
	params[id] = param;
	if next_time == nil or event_time < next_time then
		next_time = event_time;
		if _server_timer then
			_server_timer:close();
			_server_timer = nil;
		else
			_active_timers = _active_timers + 1;
		end
		_server_timer = _add_task(next_time - current_time, _on_timer);
	end
	return id;
end
local function stop(id)
	params[id] = nil;
	local result, item, result_sync = h:remove(id);
	local peek = h:peek();
	if peek ~= next_time and _server_timer then
		next_time = peek;
		_server_timer:close();
		if next_time ~= nil then
			_server_timer = _add_task(math_max(next_time - get_time(), 0), _on_timer);
		end
	end
	return result, item, result_sync;
end
local function reschedule(id, delay)
	local current_time = get_time();
	local event_time = current_time + delay;
	h:reprioritize(id, delay);
	if next_time == nil or event_time < next_time then
		next_time = event_time;
		_add_task(next_time - current_time, _on_timer);
	end
	return id;
end

return {
	add_task = add_task;
	stop = stop;
	reschedule = reschedule;
};