openstack/aodh
Code Issues Proposed changes

Use joins instead of subqueries for metadata filtering

Browse Source 
Previously, Ceilometer used subqueries to apply metadata filters. These
subqueries, however, defeated MySQL's query optimizer, leading to absurdly
bad performance. In our production setup, this patch improved performance on
two metadata filters by up to 10000x.

Change-Id: Ie3bea14d05fb63a5565e2d5cced74068955b8f08
Closes-bug: #1320772
This commit is contained in:
Koert van der Veer 2014-05-20 10:53:30 +02:00
parent 21774ae8cb
commit 578e1450d0
1 changed files with 13 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
ceilometer/storage/impl_sqlalchemy.py
View File

@ -117,20 +117,25 @@ def apply_metaquery_filter(session, query, metaquery):
:param query: Query instance
:param metaquery: dict with metadata to match on.
"""
for k, v in metaquery.iteritems():
for k, value in metaquery.iteritems():
key = k[9:] # strip out 'metadata.' prefix
try:
_model = META_TYPE_MAP[type(v)]
_model = META_TYPE_MAP[type(value)]
except KeyError:
raise NotImplementedError('Query on %(key)s is of %(value)s '
'type and is not supported' %
{"key": k, "value": type(v)})
{"key": k, "value": type(value)})
else:
meta_q = session.query(_model).\
filter(and_(_model.meta_key == key,
_model.value == v)).subquery()
query = query.filter(models.Sample.id == meta_q.c.id)
meta_alias = aliased(_model)
on_clause = and_(models.Sample.id == meta_alias.id,
meta_alias.meta_key == key)
# outer join is needed to support metaquery
# with or operator on non existent metadata field
# see: test_query_non_existing_metadata_with_result
# test case.
query = query.outerjoin(meta_alias, on_clause)
query = query.filter(meta_alias.value == value)
return query