import re import random from collections import deque from datetime import timedelta from repoze.lru import lru_cache from .timer import Timer TAG_INVALID_CHARS_RE = re.compile( r"[^\w\d_\-:/\.]", #re.UNICODE ) TAG_INVALID_CHARS_SUBS = "_" # we save and expose methods called by statsd for discovery buckets_dict = { } @lru_cache(maxsize=500) def _normalize_tags_with_cache(tag_list): return [TAG_INVALID_CHARS_RE.sub(TAG_INVALID_CHARS_SUBS, tag) for tag in tag_list] def normalize_tags(tag_list): # We have to turn our input tag list into a non-mutable tuple for it to # be hashable (and thus usable) by the @lru_cache decorator. return _normalize_tags_with_cache(tuple(tag_list)) class StatsClientBase(object): """A Base class for various statsd clients.""" def close(self): """Used to close and clean up any underlying resources.""" raise NotImplementedError() def _send(self): raise NotImplementedError() def pipeline(self): raise NotImplementedError() def timer(self, stat, rate=1, tags=None, auto_send=True): """ statsd = StatsdClient.statsd with statsd.timer('bucket_name', auto_send=True) as tmr: # This block will be timed. for i in range(0, 100000): i ** 2 # you can access time here... elapsed_ms = tmr.ms """ return Timer(self, stat, rate, tags, auto_send=auto_send) def timing(self, stat, delta, rate=1, tags=None, use_decimals=True): """ Send new timing information. `delta` can be either a number of milliseconds or a timedelta. """ if isinstance(delta, timedelta): # Convert timedelta to number of milliseconds. delta = delta.total_seconds() * 1000. if use_decimals: fmt = '%0.6f|ms' else: fmt = '%s|ms' self._send_stat(stat, fmt % delta, rate, tags) def incr(self, stat, count=1, rate=1, tags=None): """Increment a stat by `count`.""" self._send_stat(stat, '%s|c' % count, rate, tags) def decr(self, stat, count=1, rate=1, tags=None): """Decrement a stat by `count`.""" self.incr(stat, -count, rate, tags) def gauge(self, stat, value, rate=1, delta=False, tags=None): """Set a gauge value.""" if value < 0 and not delta: if rate < 1: if random.random() > rate: return with self.pipeline() as pipe: pipe._send_stat(stat, '0|g', 1) pipe._send_stat(stat, '%s|g' % value, 1) else: prefix = '+' if delta and value >= 0 else '' self._send_stat(stat, '%s%s|g' % (prefix, value), rate, tags) def set(self, stat, value, rate=1): """Set a set value.""" self._send_stat(stat, '%s|s' % value, rate) def histogram(self, stat, value, rate=1, tags=None): """Set a histogram""" self._send_stat(stat, '%s|h' % value, rate, tags) def _send_stat(self, stat, value, rate, tags=None): self._after(self._prepare(stat, value, rate, tags)) def _prepare(self, stat, value, rate, tags=None): global buckets_dict buckets_dict[stat] = 1 if rate < 1: if random.random() > rate: return value = '%s|@%s' % (value, rate) if self._prefix: stat = '%s.%s' % (self._prefix, stat) res = '%s:%s%s' % ( stat, value, ("|#" + ",".join(normalize_tags(tags))) if tags else "", ) return res def _after(self, data): if data: self._send(data) class PipelineBase(StatsClientBase): def __init__(self, client): self._client = client self._prefix = client._prefix self._stats = deque() def _send(self): raise NotImplementedError() def _after(self, data): if data is not None: self._stats.append(data) def __enter__(self): return self def __exit__(self, typ, value, tb): self.send() def send(self): if not self._stats: return self._send() def pipeline(self): return self.__class__(self)