class Memory::Leak::Monitor

We should be careful not to filter historical data, as some memory leaks may only become apparent after a long period of time. Any kind of filtering may prevent us from detecting such a leak.
A memory leak is characterised by the memory usage of the application continuing to rise over time. We can detect this by sampling memory usage and comparing it to the previous sample. If the memory usage is higher than the previous sample, we can say that the application has allocated more memory. Eventually we expect to see this stabilize, but if it continues to rise, we can say that the application has a memory leak.
Detects memory leaks by tracking heap size increases.

def as_json(...)

@returns [Hash] A serializable representation of the cluster. 
def as_json(...)
	{
		process_id: @process_id,
		current: @current,
		maximum: @maximum,
		threshold: @threshold,
		limit: @limit,
		count: @count,
	}
end

def current

@returns [Integer] The last sampled memory usage. 
def current
	@current ||= memory_usage
end

def initialize(process_id = Process.pid, maximum: nil, threshold: DEFAULT_THRESHOLD, limit: DEFAULT_LIMIT)

@parameter [Integer] The process ID to monitor.
@parameter limit [Numeric] The limit for the number of heap size increases, before we assume a memory leak.
@parameter threshold [Numeric] The threshold for heap size increases, in bytes.
@parameter maximum [Numeric] The initial maximum heap size, from which we willl track increases, in bytes.

Create a new monitor. 
def initialize(process_id = Process.pid, maximum: nil, threshold: DEFAULT_THRESHOLD, limit: DEFAULT_LIMIT)
	@process_id = process_id
	
	@maximum = maximum
	@threshold = threshold
	@limit = limit
	
	# The number of increasing heap size samples.
	@count = 0
	@current = nil
end

def leaking?

@returns [Boolean] True if a memory leak has been detected.

If the number of increasing heap size samples is greater than or equal to the limit, a memory leak is assumed.

Indicates whether a memory leak has been detected. 
def leaking?
	@count >= @limit
end

def memory_usage

@returns [Numeric] Memory usage size in bytes.

Even thought the absolute value of this number may not very useful, the relative change is useful for detecting memory leaks, and it works on most platforms.

The current resident set size (RSS) of the process. 
def memory_usage
n(["ps", "-o", "rss=", @process_id.to_s]) do |io|
 Integer(io.readlines.last) * 1024

def sample!

@yields {|sample, monitor| ...} If a memory leak is detected.

Capture a memory usage sample and yield if a memory leak is detected. 
def sample!
	@current = memory_usage
	
	if @maximum
		delta = @current - @maximum
		Console.debug(self, "Heap size captured.", current: @current, delta: delta, threshold: @threshold, maximum: @maximum)
		
		if delta > @threshold
			@maximum = @current
			@count += 1
			
			Console.debug(self, "Heap size increased.", maximum: @maximum, count: @count)
		end
	else
		Console.debug(self, "Initial heap size captured.", current: @current)
		@maximum = @current
	end
	
	return @current
end

def to_json(...)

@returns [String] The JSON representation of the cluster. 
def to_json(...)
	as_json.to_json(...)
end