openstack/cloudkitty
Code Issues Proposed changes

Optimize CloudKitty reprocessing process

Browse Source 
Currently, when a reprocessing task is scheduled, CloudKitty executes
the cleaning of the data for the reprocessing period in one hour
fashion (the default period). Therefore, for each one of the
timeframes, a delete query is sent to InfluxDB (when using it as a
backend). However, InfluxDB is not a very optimized time series database
for deletion; thus, this workflow generates quite some overhead and
slowness when reprocessing. If we clean right away the whole time
frame for the reprocessing task, and then we just reprocess it, it will
execute a single delete query in InfluxDB, which has a similar cost as
a delete to remove the data for a single time frame.

This patch optimized the reprocessing workflow to execute batch cleaning
of data in the storage backend of CloudKitty.

Change-Id: I8282f44ad837c71df0cb6c73776eafc7014ebedf
This commit is contained in:
Rafael Weingärtner 2023-07-12 19:27:04 -03:00
parent 60077a3cc4
commit 5c2f9e7f71
3 changed files with 33 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
cloudkitty/orchestrator.py
View File

@ -517,13 +517,27 @@ class ReprocessingWorker(Worker):
end_of_this_processing = tzutils.local_to_utc(end_of_this_processing)
LOG.debug("Cleaning backend [%s] data for reprocessing scope [%s] "
"for timeframe[start=%s, end=%s].",
self._storage, self.scope, timestamp, end_of_this_processing)
self._storage.delete(
begin=timestamp, end=end_of_this_processing,
filters={self.scope_key: self._tenant_id})
# If the start_reprocess_time of the reprocessing task equals to
# the current reprocessing time, it means that we have just started
# executing it. Therefore, we can clean/erase the old data in the
# reprocessing task time frame.
if tzutils.local_to_utc(self.scope.start_reprocess_time) == timestamp:
LOG.info(
"Cleaning backend [%s] data for reprocessing scope [%s] for "
"timeframe[start=%s, end=%s].", self._storage, self.scope,
self.scope.start_reprocess_time, self.scope.end_reprocess_time)
self._storage.delete(
begin=self.scope.start_reprocess_time,
end=self.scope.end_reprocess_time,
filters={self.scope_key: self._tenant_id})
else:
LOG.debug("No need to clean backend [%s] data for reprocessing "
"scope [%s] for timeframe[start=%s, end=%s]. We are "
"past the very first timestamp; therefore, the cleaning "
"for the reprocessing task period has already been "
"executed.", self._storage, self.scope,
self.scope.start_reprocess_time,
self.scope.end_reprocess_time)
LOG.debug("Executing the reprocessing of scope [%s] for "
"timeframe[start=%s, end=%s].", self.scope, timestamp,

13
cloudkitty/tests/test_orchestrator.py
View File

@ -1035,15 +1035,16 @@ class ReprocessingWorkerTest(tests.TestCase):
self, do_execute_scope_processing_mock_from_worker):
now_timestamp = tzutils.localized_now()
self.reprocessing_worker.scope.start_reprocess_time = now_timestamp
self.reprocessing_worker.do_execute_scope_processing(now_timestamp)
expected_end = tzutils.localized_now() + datetime.timedelta(
seconds=self.reprocessing_worker._period)
self.storage_mock.delete.assert_has_calls([
mock.call(begin=now_timestamp, end=expected_end,
filters={self.reprocessing_worker.scope_key:
self.reprocessing_worker._tenant_id})])
mock.call(
begin=self.reprocessing_worker.scope.start_reprocess_time,
end=self.reprocessing_worker.scope.end_reprocess_time,
filters={
self.reprocessing_worker.scope_key:
self.reprocessing_worker._tenant_id})])
do_execute_scope_processing_mock_from_worker.assert_has_calls([
mock.call(now_timestamp)])

5
releasenotes/notes/batch-delete-reprocessing-d46df15b078a42a5.yaml Normal file
View File

@ -0,0 +1,5 @@
---
features:
- |
Optimized the reprocessing workflow to execute batch cleaning
of data in the storage backend of CloudKitty.