def test_without_turbo_without_projects_needing_final(query: ClickhouseQuery) -> None:
PostReplacementConsistencyEnforcer("project_id", None).process_query(
query, HTTPRequestSettings()
)
assert query.get_condition_from_ast() == build_in("project_id", [2])
assert not query.get_from_clause().final
def test_not_many_groups_to_exclude(query: ClickhouseQuery) -> None:
state.set_config("max_group_ids_exclude", 5)
set_project_exclude_groups(2, [100, 101, 102], ReplacerState.EVENTS)
PostReplacementConsistencyEnforcer(
"project_id", ReplacerState.EVENTS
).process_query(query, HTTPRequestSettings())
assert query.get_condition_from_ast() == FunctionCall(
None,
BooleanFunctions.AND,
(
FunctionCall(
None,
"notIn",
(
FunctionCall(
None, "assumeNotNull", (Column(None, None, "group_id"),)
),
FunctionCall(
None,
"tuple",
(Literal(None, 100), Literal(None, 101), Literal(None, 102),),
),
),
),
build_in("project_id", [2]),
),
)
assert not query.get_from_clause().final
def test_not_many_groups_to_exclude(query: ClickhouseQuery) -> None:
state.set_config("max_group_ids_exclude", 5)
set_project_exclude_groups(
2,
[100, 101, 102],
ReplacerState.ERRORS,
ReplacementType.
EXCLUDE_GROUPS, # Arbitrary replacement type, no impact on tests
)
PostReplacementConsistencyEnforcer("project_id",
ReplacerState.ERRORS).process_query(
query, HTTPQuerySettings())
assert query.get_condition() == build_and(
FunctionCall(
None,
"notIn",
(
FunctionCall(None, "assumeNotNull",
(Column(None, None, "group_id"), )),
FunctionCall(
None,
"tuple",
(
Literal(None, 100),
Literal(None, 101),
Literal(None, 102),
),
),
),
),
build_in("project_id", [2]),
)
assert not query.get_from_clause().final
def test_too_many_groups_to_exclude(query: ClickhouseQuery) -> None:
state.set_config("max_group_ids_exclude", 2)
set_project_exclude_groups(2, [100, 101, 102], ReplacerState.EVENTS)
PostReplacementConsistencyEnforcer(
"project_id", ReplacerState.EVENTS
).process_query(query, HTTPRequestSettings())
assert query.get_condition_from_ast() == build_in("project_id", [2])
assert query.get_from_clause().final
def test_without_turbo_with_projects_needing_final(
query: ClickhouseQuery) -> None:
set_project_needs_final(2, ReplacerState.EVENTS)
PostReplacementConsistencyEnforcer("project_id",
ReplacerState.EVENTS).process_query(
query, HTTPRequestSettings())
assert query.get_condition_from_ast() == build_in("project_id", [2])
assert query.get_final()
def test_multiple_not_too_many_excludes(
query_with_multiple_group_ids: ClickhouseQuery, ) -> None:
"""
Query is looking for multiple groups and there are not too many groups to exclude, but
there are fewer groups queried for than replaced.
"""
enforcer = PostReplacementConsistencyEnforcer("project_id",
ReplacerState.ERRORS)
set_project_exclude_groups(
2,
[100, 101, 102],
ReplacerState.ERRORS,
ReplacementType.
EXCLUDE_GROUPS, # Arbitrary replacement type, no impact on tests
)
enforcer._set_query_final(query_with_multiple_group_ids, True)
state.set_config("max_group_ids_exclude", 5)
enforcer.process_query(query_with_multiple_group_ids, HTTPQuerySettings())
assert query_with_multiple_group_ids.get_condition() == build_and(
build_not_in("group_id", [101, 102]),
build_and(build_in("project_id", [2]),
build_in("group_id", [101, 102])),
)
assert not query_with_multiple_group_ids.get_from_clause().final
def test_single_too_many_exclude(
query_with_single_group_id: ClickhouseQuery) -> None:
"""
Query is looking for a group that has been replaced, and there are too many
groups to exclude.
"""
enforcer = PostReplacementConsistencyEnforcer("project_id",
ReplacerState.ERRORS)
set_project_exclude_groups(
2,
[100, 101, 102],
ReplacerState.ERRORS,
ReplacementType.
EXCLUDE_GROUPS, # Arbitrary replacement type, no impact on tests
)
enforcer._set_query_final(query_with_single_group_id, True)
state.set_config("max_group_ids_exclude", 2)
enforcer.process_query(query_with_single_group_id, HTTPQuerySettings())
assert query_with_single_group_id.get_condition() == build_and(
build_not_in("group_id", [101]),
build_and(build_in("project_id", [2]), build_in("group_id", [101])),
)
assert not query_with_single_group_id.get_from_clause().final
def test_without_turbo_with_projects_needing_final(
query: ClickhouseQuery) -> None:
set_project_needs_final(
2,
ReplacerState.ERRORS,
ReplacementType.
EXCLUDE_GROUPS, # Arbitrary replacement type, no impact on tests
)
PostReplacementConsistencyEnforcer("project_id",
ReplacerState.ERRORS).process_query(
query, HTTPQuerySettings())
assert query.get_condition() == build_in("project_id", [2])
assert query.get_from_clause().final
def test_too_many_groups_to_exclude(query: ClickhouseQuery) -> None:
state.set_config("max_group_ids_exclude", 2)
set_project_exclude_groups(
2,
[100, 101, 102],
ReplacerState.ERRORS,
ReplacementType.
EXCLUDE_GROUPS, # Arbitrary replacement type, no impact on tests
)
PostReplacementConsistencyEnforcer("project_id",
ReplacerState.ERRORS).process_query(
query, HTTPQuerySettings())
assert query.get_condition() == build_in("project_id", [2])
assert query.get_from_clause().final
def test_no_groups_too_many_excludes(query: ClickhouseQuery) -> None:
"""
Query has no groups, and too many to exclude.
"""
enforcer = PostReplacementConsistencyEnforcer("project_id",
ReplacerState.ERRORS)
set_project_exclude_groups(
2,
[100, 101, 102],
ReplacerState.ERRORS,
ReplacementType.
EXCLUDE_GROUPS, # Arbitrary replacement type, no impact on tests
)
enforcer._set_query_final(query, True)
state.set_config("max_group_ids_exclude", 1)
enforcer.process_query(query, HTTPQuerySettings())
assert query.get_condition() == build_in("project_id", [2])
assert query.get_from_clause().final
def test_multiple_disjoint_replaced(
query_with_multiple_group_ids: ClickhouseQuery, ) -> None:
"""
Query is looking for multiple groups and there are replaced groups, but these
sets of group ids are disjoint. (No queried groups have been replaced)
"""
enforcer = PostReplacementConsistencyEnforcer("project_id",
ReplacerState.ERRORS)
set_project_exclude_groups(
2,
[110, 120, 130],
ReplacerState.ERRORS,
ReplacementType.
EXCLUDE_GROUPS, # Arbitrary replacement type, no impact on tests
)
enforcer._set_query_final(query_with_multiple_group_ids, True)
state.set_config("max_group_ids_exclude", 5)
enforcer.process_query(query_with_multiple_group_ids, HTTPQuerySettings())
assert query_with_multiple_group_ids.get_condition() == build_and(
build_in("project_id", [2]), build_in("group_id", [101, 102]))
assert not query_with_multiple_group_ids.get_from_clause().final
def test_with_turbo(query: ClickhouseQuery) -> None:
PostReplacementConsistencyEnforcer("project_id", None).process_query(
query, HTTPRequestSettings(turbo=True)
)
assert query.get_condition_from_ast() == build_in("project_id", [2])
required_columns = [
"event_id",
"project_id",
"group_id",
"timestamp",
"deleted",
"retention_days",
]
storage = WritableTableStorage(
storage_key=StorageKey.ERRORS,
storage_set_key=StorageSetKey.EVENTS,
schema=schema,
query_processors=[
PostReplacementConsistencyEnforcer(
project_column="project_id", replacer_state_name=ReplacerState.ERRORS,
),
MappingColumnPromoter(mapping_specs={"tags": promoted_tag_columns}),
ArrayJoinKeyValueOptimizer("tags"),
PrewhereProcessor(),
],
stream_loader=KafkaStreamLoader(
processor=ErrorsProcessor(promoted_tag_columns),
default_topic="events",
replacement_topic="errors-replacements",
),
replacer_processor=ErrorsReplacer(
write_schema=schema,
read_schema=schema,
required_columns=required_columns,
tag_column_map={"tags": promoted_tag_columns, "contexts": {}},
def test_query_overlaps_replacements_processor(
query: ClickhouseQuery,
query_with_timestamp: ClickhouseQuery,
query_with_future_timestamp: ClickhouseQuery,
) -> None:
enforcer = PostReplacementConsistencyEnforcer("project_id",
ReplacerState.ERRORS)
# replacement time unknown, default to "overlaps" but no groups to exclude so shouldn't be final
enforcer._set_query_final(query_with_timestamp, True)
enforcer.process_query(query_with_timestamp, HTTPQuerySettings())
assert not query_with_timestamp.get_from_clause().final
# overlaps replacement and should be final due to too many groups to exclude
state.set_config("max_group_ids_exclude", 2)
set_project_exclude_groups(
2,
[100, 101, 102],
ReplacerState.ERRORS,
ReplacementType.
EXCLUDE_GROUPS, # Arbitrary replacement type, no impact on tests
)
enforcer._set_query_final(query_with_timestamp, False)
enforcer.process_query(query_with_timestamp, HTTPQuerySettings())
assert query_with_timestamp.get_from_clause().final
# query time range unknown and should be final due to too many groups to exclude
enforcer._set_query_final(query, False)
enforcer.process_query(query, HTTPQuerySettings())
assert query.get_from_clause().final
# doesn't overlap replacements
enforcer._set_query_final(query_with_future_timestamp, True)
enforcer.process_query(query_with_future_timestamp, HTTPQuerySettings())
assert not query_with_future_timestamp.get_from_clause().final
prewhere_candidates = [
"event_id",
"group_id",
"tags[sentry:release]",
"sentry:release",
"message",
"title",
"environment",
"project_id",
]
query_processors = [
PostReplacementConsistencyEnforcer(
project_column="project_id",
# key migration is on going. As soon as all the keys we are interested
# into in redis are stored with "EVENTS" in the name, we can change this.
replacer_state_name=None,
),
EventsColumnProcessor(),
MappingColumnPromoter(
mapping_specs={
"tags": ChainMap(
{col.flattened: col.flattened for col in promoted_tag_columns},
get_promoted_context_tag_col_mapping(),
),
"contexts": get_promoted_context_col_mapping(),
},
),
# This processor must not be ported to the errors dataset. We should
# not support promoting tags/contexts with boolean values. There is
# no way to convert them back consistently to the value provided by
