8000 fix usage analytics to properly register and not count if disabled by bentsku · Pull Request #11848 · localstack/localstack · GitHub
[go: up one dir, main page]
Skip to content

fix usage analytics to properly register and not count if disabled #11848

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.

Sign up for GitHub

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

Merged
merged 6 commits into from
Nov 18, 2024
Merged

fix usage analytics to properly register and not count if disabled #11848

Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
a724c10
fix usage analytics to properly register and not count if disabled
bentsku Nov 14, 2024
f04328b
use default dict instead of list to reduce memory
bentsku Nov 15, 2024
4df164b
use itertools.count
bentsku Nov 15, 2024
533d9dd
remove conditional loading
bentsku Nov 18, 2024
27c7c52
fix counting start
bentsku Nov 18, 2024
41fd489
oopsie
bentsku Nov 18, 2024
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
42 changes: 27 additions & 15 deletions localstack-core/localstack/utils/analytics/usage.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,8 +1,11 @@
import datetime
import math
from collections import defaultdict
from itertools import count
from typing import Any

from localstack import config
from localstack.runtime import hooks
from localstack.utils.analytics import get_session_id
from localstack.utils.analytics.events import Event, EventMetadata
from localstack.utils.analytics.publisher import AnalyticsClientPublisher
Expand All @@ -11,6 +14,8 @@
collector_registry: dict[str, Any] = dict()

# TODO: introduce some base abstraction for the counters after gather some initial experience working with it
# we could probably do intermediate aggregations over time to avoid unbounded counters for very long LS sessions
# for now, we can recommend to use config.DISABLE_EVENTS=1


class UsageSetCounter:
Expand All @@ -25,23 +30,22 @@ class UsageSetCounter:
my_feature_counter.aggregate() # returns {"python3.7": 1, "nodejs16.x": 2}
"""

state: list[str]
state: dict[str, int]
_counter: dict[str, count]
namespace: str

def __init__(self, namespace: str):
self.state = list()
self.enabled = not config.DISABLE_EVENTS
self.state = {}
self._counter = defaultdict(lambda: count(1))
self.namespace = namespace
collector_registry[namespace] = self

def record(self, value: str):
self.state.append(value)
if self.enabled:
self.state[value] = next(self._counter[value])

def aggregate(self) -> dict:
result = {}
for a in self.state:
result.setdefault(a, 0)
result[a] = result[a] + 1
return result
return self.state


class UsageCounter:
Expand All @@ -62,21 +66,26 @@ class UsageCounter:
aggregations: list[str]

def __init__(self, namespace: str, aggregations: list[str]):
self.state = list()
self.enabled = not config.DISABLE_EVENTS
self.state = []
self.namespace = namespace
self.aggregations = aggregations
collector_registry[namespace] = self

def increment(self):
self.state.append(1)
# TODO: we should instead have different underlying datastructures to store the state, and have no-op operations
# when config.DISABLE_EVENTS is set
if self.enabled:
self.state.append(1)

def record_value(self, value: int | float):
self.state.append(value)
if self.enabled:
self.state.append(value)

def aggregate(self) -> dict:
result = {}
for aggregation in self.aggregations:
if self.state:
if self.state:
for aggregation in self.aggregations:
match aggregation:
case "sum":
result[aggregation] = sum(self.state)
Expand All @@ -88,7 +97,9 @@ def aggregate(self) -> dict:
result[aggregation] = sum(self.state) / len(self.state)
case "median":
median_index = math.floor(len(self.state) / 2)
result[aggregation] = self.state[median_index]
result[aggregation] = sorted(self.state)[median_index]
case "count":
result[aggregation] = len(self.state)
case _:
raise Exception(f"Unsupported aggregation: {aggregation}")
return result
Expand All @@ -101,6 +112,7 @@ def aggregate() -> dict:
return aggregated_payload


@hooks.on_infra_shutdown()
def aggregate_and_send():
"""
Aggregates data from all registered usage trackers and immediately sends the aggregated result to the analytics service.
Expand Down
Loading
0