synapse-old/synapse/storage/databases/main/monthly_active_users.py

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

439 lines
17 KiB
Python
Raw Normal View History

2018-08-03 06:49:53 -06:00
# Copyright 2018 New Vector
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import logging
from typing import TYPE_CHECKING, Dict, List, Optional, Tuple, cast
2018-08-03 06:49:53 -06:00
from synapse.metrics.background_process_metrics import wrap_as_background_process
from synapse.storage.database import (
DatabasePool,
LoggingDatabaseConnection,
LoggingTransaction,
make_in_list_sql_clause,
)
from synapse.storage.databases.main.registration import RegistrationWorkerStore
from synapse.util.caches.descriptors import cached
from synapse.util.threepids import canonicalise_email
2018-07-31 09:36:24 -06:00
if TYPE_CHECKING:
from synapse.server import HomeServer
logger = logging.getLogger(__name__)
2018-08-06 16:25:25 -06:00
2018-08-03 10:55:50 -06:00
# Number of msec of granularity to store the monthly_active_user timestamp
# This means it is not necessary to update the table on every request
LAST_SEEN_GRANULARITY = 60 * 60 * 1000
2018-07-31 09:36:24 -06:00
class MonthlyActiveUsersWorkerStore(RegistrationWorkerStore):
def __init__(
self,
database: DatabasePool,
db_conn: LoggingDatabaseConnection,
hs: "HomeServer",
):
super().__init__(database, db_conn, hs)
2018-07-31 09:36:24 -06:00
self._clock = hs.get_clock()
2018-08-03 06:49:53 -06:00
self.hs = hs
if hs.config.redis.redis_enabled:
# If we're using Redis, we can shift this update process off to
# the background worker
self._update_on_this_worker = hs.config.worker.run_background_tasks
else:
# If we're NOT using Redis, this must be handled by the master
self._update_on_this_worker = hs.get_instance_name() == "master"
self._limit_usage_by_mau = hs.config.server.limit_usage_by_mau
self._max_mau_value = hs.config.server.max_mau_value
self._mau_stats_only = hs.config.server.mau_stats_only
if self._update_on_this_worker:
# Do not add more reserved users than the total allowable number
self.db_pool.new_transaction(
db_conn,
"initialise_mau_threepids",
[],
[],
[],
self._initialise_reserved_users,
hs.config.server.mau_limits_reserved_threepids[: self._max_mau_value],
)
@cached(num_args=0)
async def get_monthly_active_count(self) -> int:
"""Generates current count of monthly active users
Returns:
Number of current monthly active users
"""
def _count_users(txn: LoggingTransaction) -> int:
# Exclude app service users
sql = """
SELECT COUNT(*)
FROM monthly_active_users
LEFT JOIN users
ON monthly_active_users.user_id=users.name
WHERE (users.appservice_id IS NULL OR users.appservice_id = '');
"""
txn.execute(sql)
(count,) = cast(Tuple[int], txn.fetchone())
return count
return await self.db_pool.runInteraction("count_users", _count_users)
@cached(num_args=0)
async def get_monthly_active_count_by_service(self) -> Dict[str, int]:
"""Generates current count of monthly active users broken down by service.
A service is typically an appservice but also includes native matrix users.
Since the `monthly_active_users` table is populated from the `user_ips` table
`config.appservice.track_appservice_user_ips` must be set to `true` for this
method to return anything other than native matrix users.
Returns:
A mapping between app_service_id and the number of occurrences.
"""
def _count_users_by_service(txn: LoggingTransaction) -> Dict[str, int]:
sql = """
SELECT COALESCE(appservice_id, 'native'), COUNT(*)
FROM monthly_active_users
LEFT JOIN users ON monthly_active_users.user_id=users.name
GROUP BY appservice_id;
"""
txn.execute(sql)
result = cast(List[Tuple[str, int]], txn.fetchall())
return dict(result)
return await self.db_pool.runInteraction(
"count_users_by_service", _count_users_by_service
)
async def get_monthly_active_users_by_service(
self, start_timestamp: Optional[int] = None, end_timestamp: Optional[int] = None
) -> List[Tuple[str, str]]:
"""Generates list of monthly active users and their services.
Please see "get_monthly_active_count_by_service" docstring for more details
about services.
Arguments:
start_timestamp: If specified, only include users that were first active
at or after this point
end_timestamp: If specified, only include users that were first active
at or before this point
Returns:
A list of tuples (appservice_id, user_id). "native" is emitted as the
appservice for users that don't come from appservices (i.e. native Matrix
users).
"""
if start_timestamp is not None and end_timestamp is not None:
where_clause = 'WHERE "timestamp" >= ? and "timestamp" <= ?'
query_params = [start_timestamp, end_timestamp]
elif start_timestamp is not None:
where_clause = 'WHERE "timestamp" >= ?'
query_params = [start_timestamp]
elif end_timestamp is not None:
where_clause = 'WHERE "timestamp" <= ?'
query_params = [end_timestamp]
else:
where_clause = ""
query_params = []
def _list_users(txn: LoggingTransaction) -> List[Tuple[str, str]]:
sql = f"""
SELECT COALESCE(appservice_id, 'native'), user_id
FROM monthly_active_users
LEFT JOIN users ON monthly_active_users.user_id=users.name
{where_clause};
"""
txn.execute(sql, query_params)
return cast(List[Tuple[str, str]], txn.fetchall())
return await self.db_pool.runInteraction("list_users", _list_users)
async def get_registered_reserved_users(self) -> List[str]:
"""Of the reserved threepids defined in config, retrieve those that are associated
with registered users
Returns:
User IDs of actual users that are reserved
"""
users = []
for tp in self.hs.config.server.mau_limits_reserved_threepids[
: self.hs.config.server.max_mau_value
]:
user_id = await self.hs.get_datastores().main.get_user_id_by_threepid(
tp["medium"], canonicalise_email(tp["address"])
)
if user_id:
users.append(user_id)
return users
@cached(num_args=1)
async def user_last_seen_monthly_active(self, user_id: str) -> Optional[int]:
"""
Checks if a given user is part of the monthly active user group
Arguments:
user_id: user to add/update
Return:
Timestamp since last seen, None if never seen
"""
return await self.db_pool.simple_select_one_onecol(
table="monthly_active_users",
keyvalues={"user_id": user_id},
retcol="timestamp",
allow_none=True,
desc="user_last_seen_monthly_active",
)
@wrap_as_background_process("reap_monthly_active_users")
async def reap_monthly_active_users(self) -> None:
"""Cleans out monthly active user table to ensure that no stale
2018-07-31 09:36:24 -06:00
entries exist.
"""
def _reap_users(txn: LoggingTransaction, reserved_users: List[str]) -> None:
"""
Args:
reserved_users (tuple): reserved users to preserve
"""
2018-08-03 06:49:53 -06:00
thirty_days_ago = int(self._clock.time_msec()) - (1000 * 60 * 60 * 24 * 30)
in_clause, in_clause_args = make_in_list_sql_clause(
self.database_engine, "user_id", reserved_users
)
txn.execute(
"DELETE FROM monthly_active_users WHERE timestamp < ? AND NOT %s"
% (in_clause,),
[thirty_days_ago] + in_clause_args,
)
2018-08-06 14:55:54 -06:00
if self._limit_usage_by_mau:
# If MAU user count still exceeds the MAU threshold, then delete on
# a least recently active basis.
# Note it is not possible to write this query using OFFSET due to
# incompatibilities in how sqlite and postgres support the feature.
# Sqlite requires 'LIMIT -1 OFFSET ?', the LIMIT must be present,
# while Postgres does not require 'LIMIT', but also does not support
# negative LIMIT values. So there is no way to write it that both can
# support
# Limit must be >= 0 for postgres
num_of_non_reserved_users_to_remove = max(
self._max_mau_value - len(reserved_users), 0
)
# It is important to filter reserved users twice to guard
# against the case where the reserved user is present in the
# SELECT, meaning that a legitimate mau is deleted.
sql = """
DELETE FROM monthly_active_users
WHERE user_id NOT IN (
SELECT user_id FROM monthly_active_users
WHERE NOT %s
ORDER BY timestamp DESC
LIMIT ?
)
AND NOT %s
""" % (
in_clause,
in_clause,
)
query_args = (
in_clause_args
+ [num_of_non_reserved_users_to_remove]
+ in_clause_args
)
txn.execute(sql, query_args)
# It seems poor to invalidate the whole cache. Postgres supports
# 'Returning' which would allow me to invalidate only the
# specific users, but sqlite has no way to do this and instead
# I would need to SELECT and the DELETE which without locking
# is racy.
# Have resolved to invalidate the whole cache for now and do
# something about it if and when the perf becomes significant
self._invalidate_all_cache_and_stream(
txn, self.user_last_seen_monthly_active
)
self._invalidate_cache_and_stream(txn, self.get_monthly_active_count, ())
reserved_users = await self.get_registered_reserved_users()
await self.db_pool.runInteraction(
"reap_monthly_active_users", _reap_users, reserved_users
)
2018-09-12 04:58:52 -06:00
def _initialise_reserved_users(
self, txn: LoggingTransaction, threepids: List[dict]
) -> None:
"""Ensures that reserved threepids are accounted for in the MAU table, should
be called on start up.
Args:
txn:
threepids: List of threepid dicts to reserve
"""
assert (
self._update_on_this_worker
), "This worker is not designated to update MAUs"
# XXX what is this function trying to achieve? It upserts into
# monthly_active_users for each *registered* reserved mau user, but why?
#
# - shouldn't there already be an entry for each reserved user (at least
# if they have been active recently)?
#
# - if it's important that the timestamp is kept up to date, why do we only
# run this at startup?
for tp in threepids:
user_id = self.get_user_id_by_threepid_txn(txn, tp["medium"], tp["address"])
if user_id:
is_support = self.is_support_user_txn(txn, user_id)
if not is_support:
# We do this manually here to avoid hitting #6791
self.db_pool.simple_upsert_txn(
txn,
table="monthly_active_users",
keyvalues={"user_id": user_id},
values={"timestamp": int(self._clock.time_msec())},
)
else:
logger.warning("mau limit reserved threepid %s not found in db" % tp)
async def upsert_monthly_active_user(self, user_id: str) -> None:
2018-10-24 10:18:08 -06:00
"""Updates or inserts the user into the monthly active user table, which
is used to track the current MAU usage of the server
Args:
user_id: user to add/update
"""
assert (
self._update_on_this_worker
), "This worker is not designated to update MAUs"
# Support user never to be included in MAU stats. Note I can't easily call this
# from upsert_monthly_active_user_txn because then I need a _txn form of
# is_support_user which is complicated because I want to cache the result.
# Therefore I call it here and ignore the case where
# upsert_monthly_active_user_txn is called directly from
# _initialise_reserved_users reasoning that it would be very strange to
# include a support user in this context.
is_support = await self.is_support_user(user_id)
if is_support:
return
await self.db_pool.runInteraction(
"upsert_monthly_active_user", self.upsert_monthly_active_user_txn, user_id
)
2018-10-25 07:58:59 -06:00
def upsert_monthly_active_user_txn(
self, txn: LoggingTransaction, user_id: str
) -> None:
2018-10-24 10:18:08 -06:00
"""Updates or inserts monthly active user member
2018-10-25 09:13:43 -06:00
We consciously do not call is_support_txn from this method because it
is not possible to cache the response. is_support_txn will be false in
almost all cases, so it seems reasonable to call it only for
upsert_monthly_active_user and to call is_support_txn manually
for cases where upsert_monthly_active_user_txn is called directly,
like _initialise_reserved_users
In short, don't call this method with support users. (Support users
should not appear in the MAU stats).
2018-10-24 10:18:08 -06:00
Args:
txn (cursor):
user_id (str): user to add/update
2018-07-31 09:36:24 -06:00
"""
assert (
self._update_on_this_worker
), "This worker is not designated to update MAUs"
2018-09-26 09:16:41 -06:00
# Am consciously deciding to lock the table on the basis that is ought
# never be a big table and alternative approaches (batching multiple
# upserts into a single txn) introduced a lot of extra complexity.
# See https://github.com/matrix-org/synapse/issues/3854 for more
self.db_pool.simple_upsert_txn(
txn,
2018-07-31 09:36:24 -06:00
table="monthly_active_users",
keyvalues={"user_id": user_id},
values={"timestamp": int(self._clock.time_msec())},
2018-07-31 09:36:24 -06:00
)
2018-10-24 10:18:08 -06:00
self._invalidate_cache_and_stream(txn, self.get_monthly_active_count, ())
self._invalidate_cache_and_stream(
txn, self.get_monthly_active_count_by_service, ()
)
self._invalidate_cache_and_stream(
txn, self.user_last_seen_monthly_active, (user_id,)
)
2018-08-03 10:55:50 -06:00
async def populate_monthly_active_users(self, user_id: str) -> None:
2018-08-03 10:55:50 -06:00
"""Checks on the state of monthly active user limits and optionally
add the user to the monthly active tables
Args:
user_id(str): the user_id to query
"""
assert (
self._update_on_this_worker
), "This worker is not designated to update MAUs"
if self._limit_usage_by_mau or self._mau_stats_only:
2018-09-06 10:22:53 -06:00
# Trial users and guests should not be included as part of MAU group
is_guest = await self.is_guest(user_id)
if is_guest:
return
is_trial = await self.is_trial_user(user_id)
2018-08-23 12:17:08 -06:00
if is_trial:
return
last_seen_timestamp = await self.user_last_seen_monthly_active(user_id)
2018-08-03 10:55:50 -06:00
now = self.hs.get_clock().time_msec()
# We want to reduce to the total number of db writes, and are happy
# to trade accuracy of timestamp in order to lighten load. This means
# We always insert new users (where MAU threshold has not been reached),
# but only update if we have not previously seen the user for
# LAST_SEEN_GRANULARITY ms
2018-08-03 10:55:50 -06:00
if last_seen_timestamp is None:
# In the case where mau_stats_only is True and limit_usage_by_mau is
# False, there is no point in checking get_monthly_active_count - it
# adds no value and will break the logic if max_mau_value is exceeded.
if not self._limit_usage_by_mau:
await self.upsert_monthly_active_user(user_id)
else:
count = await self.get_monthly_active_count()
if count < self._max_mau_value:
await self.upsert_monthly_active_user(user_id)
2018-08-03 10:55:50 -06:00
elif now - last_seen_timestamp > LAST_SEEN_GRANULARITY:
await self.upsert_monthly_active_user(user_id)