Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add sum, livemin, and livemax multiprocess modes for Gauges #794

Merged
merged 3 commits into from
Apr 25, 2022
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 9 additions & 6 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -609,14 +609,17 @@ def child_exit(server, worker):

**4. Metrics tuning (Gauge)**:

When `Gauge` metrics are used, additional tuning needs to be performed.
When `Gauge`s are used in multiprocess applications,
you must decide how to handle the metrics reported by each process.
Gauges have several modes they can run in, which can be selected with the `multiprocess_mode` parameter.

- 'all': Default. Return a timeseries per process alive or dead.
- 'liveall': Return a timeseries per process that is still alive.
- 'livesum': Return a single timeseries that is the sum of the values of alive processes.
- 'max': Return a single timeseries that is the maximum of the values of all processes, alive or dead.
- 'min': Return a single timeseries that is the minimum of the values of all processes, alive or dead.
- 'all': Default. Return a timeseries per process (alive or dead), labelled by the process's `pid` (the label is added internally).
- 'min': Return a single timeseries that is the minimum of the values of all processes (alive or dead).
- 'max': Return a single timeseries that is the maximum of the values of all processes (alive or dead).
- 'sum': Return a single timeseries that is the sum of the values of all processes (alive or dead).

Prepend 'live' to the beginning of the mode to return the same result but only considering living processes
(e.g., 'liveall, 'livesum', 'livemax', 'livemin').

```python
from prometheus_client import Gauge
Expand Down
2 changes: 1 addition & 1 deletion prometheus_client/metrics.py
Original file line number Diff line number Diff line change
Expand Up @@ -346,7 +346,7 @@ def f():
d.set_function(lambda: len(my_dict))
"""
_type = 'gauge'
_MULTIPROC_MODES = frozenset(('min', 'max', 'livesum', 'liveall', 'all'))
_MULTIPROC_MODES = frozenset(('all', 'liveall', 'min', 'livemin', 'max', 'livemax', 'sum', 'livesum'))

def __init__(self,
name: str,
Expand Down
21 changes: 12 additions & 9 deletions prometheus_client/multiprocess.py
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,7 @@
import os
import warnings

from .metrics import Gauge
from .metrics_core import Metric
from .mmap_dict import MmapedDict
from .samples import Sample
Expand Down Expand Up @@ -63,8 +64,8 @@ def _parse_key(key):
try:
file_values = MmapedDict.read_all_values_from_file(f)
except FileNotFoundError:
if typ == 'gauge' and parts[1] in ('liveall', 'livesum'):
# Those files can disappear between the glob of collect
if typ == 'gauge' and parts[1].startswith('live'):
# Files for 'live*' gauges can be deleted between the glob of collect
# and now (via a mark_process_dead call) so don't fail if
# the file is missing
continue
Expand Down Expand Up @@ -96,15 +97,15 @@ def _accumulate_metrics(metrics, accumulate):
name, labels, value, timestamp, exemplar = s
if metric.type == 'gauge':
without_pid_key = (name, tuple(l for l in labels if l[0] != 'pid'))
if metric._multiprocess_mode == 'min':
if metric._multiprocess_mode in ('min', 'livemin'):
current = samples_setdefault(without_pid_key, value)
if value < current:
samples[without_pid_key] = value
elif metric._multiprocess_mode == 'max':
elif metric._multiprocess_mode in ('max', 'livemax'):
current = samples_setdefault(without_pid_key, value)
if value > current:
samples[without_pid_key] = value
elif metric._multiprocess_mode == 'livesum':
elif metric._multiprocess_mode in ('sum', 'livesum'):
samples[without_pid_key] += value
else: # all/liveall
samples[(name, labels)] = value
Expand Down Expand Up @@ -152,11 +153,13 @@ def collect(self):
return self.merge(files, accumulate=True)


_LIVE_GAUGE_MULTIPROCESS_MODES = {m for m in Gauge._MULTIPROC_MODES if m.startswith('live')}


def mark_process_dead(pid, path=None):
"""Do bookkeeping for when one process dies in a multi-process setup."""
if path is None:
path = os.environ.get('PROMETHEUS_MULTIPROC_DIR', os.environ.get('prometheus_multiproc_dir'))
for f in glob.glob(os.path.join(path, f'gauge_livesum_{pid}.db')):
os.remove(f)
for f in glob.glob(os.path.join(path, f'gauge_liveall_{pid}.db')):
os.remove(f)
for mode in _LIVE_GAUGE_MULTIPROCESS_MODES:
for f in glob.glob(os.path.join(path, f'gauge_{mode}_{pid}.db')):
os.remove(f)
33 changes: 33 additions & 0 deletions tests/test_multiprocess.py
Original file line number Diff line number Diff line change
Expand Up @@ -132,6 +132,17 @@ def test_gauge_min(self):
g2.set(2)
self.assertEqual(1, self.registry.get_sample_value('g'))

def test_gauge_livemin(self):
g1 = Gauge('g', 'help', registry=None, multiprocess_mode='livemin')
values.ValueClass = MultiProcessValue(lambda: 456)
g2 = Gauge('g', 'help', registry=None, multiprocess_mode='livemin')
self.assertEqual(0, self.registry.get_sample_value('g'))
g1.set(1)
g2.set(2)
self.assertEqual(1, self.registry.get_sample_value('g'))
mark_process_dead(123, os.environ['PROMETHEUS_MULTIPROC_DIR'])
self.assertEqual(2, self.registry.get_sample_value('g'))

def test_gauge_max(self):
g1 = Gauge('g', 'help', registry=None, multiprocess_mode='max')
values.ValueClass = MultiProcessValue(lambda: 456)
Expand All @@ -141,6 +152,28 @@ def test_gauge_max(self):
g2.set(2)
self.assertEqual(2, self.registry.get_sample_value('g'))

def test_gauge_livemax(self):
g1 = Gauge('g', 'help', registry=None, multiprocess_mode='livemax')
values.ValueClass = MultiProcessValue(lambda: 456)
g2 = Gauge('g', 'help', registry=None, multiprocess_mode='livemax')
self.assertEqual(0, self.registry.get_sample_value('g'))
g1.set(2)
g2.set(1)
self.assertEqual(2, self.registry.get_sample_value('g'))
mark_process_dead(123, os.environ['PROMETHEUS_MULTIPROC_DIR'])
self.assertEqual(1, self.registry.get_sample_value('g'))

def test_gauge_sum(self):
g1 = Gauge('g', 'help', registry=None, multiprocess_mode='sum')
values.ValueClass = MultiProcessValue(lambda: 456)
g2 = Gauge('g', 'help', registry=None, multiprocess_mode='sum')
self.assertEqual(0, self.registry.get_sample_value('g'))
g1.set(1)
g2.set(2)
self.assertEqual(3, self.registry.get_sample_value('g'))
mark_process_dead(123, os.environ['PROMETHEUS_MULTIPROC_DIR'])
self.assertEqual(3, self.registry.get_sample_value('g'))

def test_gauge_livesum(self):
g1 = Gauge('g', 'help', registry=None, multiprocess_mode='livesum')
values.ValueClass = MultiProcessValue(lambda: 456)
Expand Down