def __init__(self) -> None:
storage = get_writable_storage(StorageKey.SPANS)
super().__init__(
storages=[storage],
query_pipeline_builder=SimplePipelineBuilder(
query_plan_builder=SingleStorageQueryPlanBuilder(
storage=storage,
mappers=TranslationMappers(
subscriptables=[
SubscriptableMapper(None, "tags", None, "tags")
],
),
),
),
abstract_column_set=ColumnSet(
[
("project_id", UInt(64)),
("transaction_id", UUID()),
("trace_id", UUID()),
("transaction_span_id", UInt(64)),
("span_id", UInt(64)),
("parent_span_id", UInt(64, Modifiers(nullable=True))),
("transaction_name", String()),
("op", String()),
("status", UInt(8)),
("start_ts", DateTime()),
("start_ns", UInt(32)),
("finish_ts", DateTime()),
("finish_ns", UInt(32)),
("duration_ms", UInt(32)),
("tags", Nested([("key", String()), ("value", String())])),
]
),
join_relationships={
"contained": JoinRelationship(
rhs_entity=EntityKey.TRANSACTIONS,
columns=[
("project_id", "project_id"),
("transaction_span_id", "span_id"),
],
join_type=JoinType.INNER,
equivalences=[
ColumnEquivalence("transaction_id", "event_id"),
ColumnEquivalence("transaction_name", "transaction_name"),
ColumnEquivalence("trace_id", "trace_id"),
],
)
},
writable_storage=storage,
validators=[EntityRequiredColumnValidator({"project_id"})],
required_time_column=None,
)
def __init__(self) -> None:
read_columns = ColumnSet([
("org_id", UInt(64)),
("project_id", UInt(64)),
("key_id", Nullable(UInt(64))),
("timestamp", DateTime()),
("outcome", UInt(8)),
("reason", LowCardinality(Nullable(String()))),
("event_id", Nullable(UUID())),
])
read_schema = MergeTreeSchema(
columns=read_columns,
local_table_name="outcomes_raw_local",
dist_table_name="outcomes_raw_dist",
order_by="(org_id, project_id, timestamp)",
partition_by="(toMonday(timestamp))",
settings={"index_granularity": 16384},
migration_function=outcomes_raw_migrations,
)
dataset_schemas = DatasetSchemas(read_schema=read_schema,
write_schema=None,
intermediary_schemas=[])
super().__init__(
dataset_schemas=dataset_schemas,
time_group_columns={"time": "timestamp"},
time_parse_columns=("timestamp", ),
)
def __init__(self):
read_columns = ColumnSet([
('org_id', UInt(64)),
('project_id', UInt(64)),
('key_id', Nullable(UInt(64))),
('timestamp', DateTime()),
('outcome', UInt(8)),
('reason', LowCardinality(Nullable(String()))),
('event_id', Nullable(UUID())),
])
read_schema = MergeTreeSchema(
columns=read_columns,
local_table_name='outcomes_raw_local',
dist_table_name='outcomes_raw_dist',
order_by='(org_id, project_id, timestamp)',
partition_by='(toMonday(timestamp))',
settings={'index_granularity': 16384})
dataset_schemas = DatasetSchemas(read_schema=read_schema,
write_schema=None,
intermediary_schemas=[])
super().__init__(dataset_schemas=dataset_schemas,
time_group_columns={
'time': 'timestamp',
},
time_parse_columns=('timestamp', ))
def __backward_migrations(
self, table_name: str) -> Sequence[operations.SqlOperation]:
return [
operations.ModifyColumn(
storage_set=StorageSetKey.TRANSACTIONS,
table_name=table_name,
column=Column("trace_id", UUID()),
),
]
def __forward_migrations(
self, table_name: str) -> Sequence[operations.SqlOperation]:
return [
operations.AddColumn(
storage_set=StorageSetKey.DISCOVER,
table_name=table_name,
column=Column("trace_id", UUID(Modifiers(nullable=True))),
after="contexts",
),
]
def forwards_dist(self) -> Sequence[operations.SqlOperation]:
return [
operations.AddColumn(
storage_set=StorageSetKey.EVENTS,
table_name="errors_dist",
column=Column("hierarchical_hashes", Array(UUID())),
after="primary_hash",
),
operations.AddColumn(
storage_set=StorageSetKey.EVENTS,
table_name="sentry_dist",
column=Column("hierarchical_hashes", Array(FixedString(32)),),
after="primary_hash",
),
]
def __init__(self) -> None:
columns = ColumnSet(
[
("project_id", UInt(64)),
("event_id", UUID()),
("trace_id", UUID()),
("span_id", UInt(64)),
("transaction_name", LowCardinality(String())),
(
"transaction_hash",
Materialized(UInt(64), "cityHash64(transaction_name)",),
),
("transaction_op", LowCardinality(String())),
("transaction_status", WithDefault(UInt(8), UNKNOWN_SPAN_STATUS)),
("start_ts", DateTime()),
("start_ms", UInt(16)),
("_start_date", Materialized(Date(), "toDate(start_ts)"),),
("finish_ts", DateTime()),
("finish_ms", UInt(16)),
("_finish_date", Materialized(Date(), "toDate(finish_ts)"),),
("duration", UInt(32)),
("platform", LowCardinality(String())),
("environment", LowCardinality(Nullable(String()))),
("release", LowCardinality(Nullable(String()))),
("dist", LowCardinality(Nullable(String()))),
("ip_address_v4", Nullable(IPv4())),
("ip_address_v6", Nullable(IPv6())),
("user", WithDefault(String(), "''",)),
("user_hash", Materialized(UInt(64), "cityHash64(user)"),),
("user_id", Nullable(String())),
("user_name", Nullable(String())),
("user_email", Nullable(String())),
("sdk_name", WithDefault(LowCardinality(String()), "''")),
("sdk_version", WithDefault(LowCardinality(String()), "''")),
("tags", Nested([("key", String()), ("value", String())])),
("_tags_flattened", String()),
("contexts", Nested([("key", String()), ("value", String())])),
("_contexts_flattened", String()),
("partition", UInt(16)),
("offset", UInt(64)),
("retention_days", UInt(16)),
("deleted", UInt(8)),
]
)
schema = ReplacingMergeTreeSchema(
columns=columns,
local_table_name="transactions_local",
dist_table_name="transactions_dist",
mandatory_conditions=[],
prewhere_candidates=["event_id", "project_id"],
order_by="(project_id, _finish_date, transaction_name, cityHash64(span_id))",
partition_by="(retention_days, toMonday(_finish_date))",
version_column="deleted",
sample_expr=None,
migration_function=transactions_migrations,
)
dataset_schemas = DatasetSchemas(read_schema=schema, write_schema=schema,)
self.__tags_processor = TagColumnProcessor(
columns=columns,
promoted_columns=self._get_promoted_columns(),
column_tag_map=self._get_column_tag_map(),
)
super().__init__(
dataset_schemas=dataset_schemas,
table_writer=TransactionsTableWriter(
write_schema=schema,
stream_loader=KafkaStreamLoader(
processor=TransactionsMessageProcessor(), default_topic="events",
),
),
time_group_columns={
"bucketed_start": "start_ts",
"bucketed_end": "finish_ts",
},
time_parse_columns=("start_ts", "finish_ts"),
)
from abc import ABC
from typing import Mapping, Sequence
from unittest.mock import Mock
from snuba.clickhouse.columns import UUID, ColumnSet, String, UInt
from snuba.datasets.entities import EntityKey
from snuba.datasets.entity import Entity
from snuba.query.data_source.join import ColumnEquivalence, JoinRelationship, JoinType
from snuba.query.extensions import QueryExtension
from snuba.query.processors import QueryProcessor
EVENTS_SCHEMA = ColumnSet([
("event_id", UUID()),
("project_id", UInt(32)),
("message", String()),
("group_id", UInt(32)),
("user_id", UInt(64)),
])
GROUPS_SCHEMA = ColumnSet([
("id", UUID()),
("project_id", UInt(32)),
("message", String()),
("user_id", UInt(64)),
])
GROUPS_ASSIGNEE = ColumnSet([
("group_id", UUID()),
("project_id", UInt(32)),
("message", String()),
("user_id", UInt(64)),
Nullable,
String,
UInt,
UUID,
)
from snuba.datasets.dataset_schemas import StorageSchemas
from snuba.datasets.querylog_processor import QuerylogProcessor
from snuba.datasets.schemas.tables import MergeTreeSchema
from snuba.datasets.storage import WritableTableStorage
from snuba.datasets.table_storage import TableWriter, KafkaStreamLoader
status_type = Enum([("success", 0), ("error", 1), ("rate-limited", 2)])
columns = ColumnSet([
("request_id", UUID()),
("request_body", String()),
("referrer", LowCardinality(String())),
("dataset", LowCardinality(String())),
("projects", Array(UInt(64))),
("organization", Nullable(UInt(64))),
("timestamp", DateTime()),
("duration_ms", UInt(32)),
("status", status_type),
(
"clickhouse_queries",
Nested([
("sql", String()),
("status", status_type),
("trace_id", Nullable(UUID())),
("duration_ms", UInt(32)),
from typing import Sequence
from snuba.clickhouse.columns import UUID, Column, DateTime, String, UInt
from snuba.clusters.storage_sets import StorageSetKey
from snuba.migrations import migration, operations, table_engines
from snuba.migrations.columns import MigrationModifiers as Modifiers
raw_columns = [
Column("org_id", UInt(64)),
Column("project_id", UInt(64)),
Column("key_id", UInt(64, Modifiers(nullable=True))),
Column("timestamp", DateTime()),
Column("outcome", UInt(8)),
Column("reason", String(Modifiers(nullable=True, low_cardinality=True))),
Column("event_id", UUID(Modifiers(nullable=True))),
]
hourly_columns = [
Column("org_id", UInt(64)),
Column("project_id", UInt(64)),
Column("key_id", UInt(64)),
Column("timestamp", DateTime()),
Column("outcome", UInt(8)),
Column("reason", String(Modifiers(low_cardinality=True))),
Column("times_seen", UInt(64)),
]
materialized_view_columns = [
Column("org_id", UInt(64)),
Column("project_id", UInt(64)),
Column("key_id", UInt(64)),
from snuba.datasets.storages import StorageKey
from snuba.datasets.table_storage import build_kafka_stream_loader_from_settings
from snuba.query.processors.prewhere import PrewhereProcessor
WRITE_LOCAL_TABLE_NAME = "sessions_raw_local"
WRITE_DIST_TABLE_NAME = "sessions_raw_dist"
READ_LOCAL_TABLE_NAME = "sessions_hourly_local"
READ_DIST_TABLE_NAME = "sessions_hourly_dist"
READ_LOCAL_MV_NAME = "sessions_hourly_mv_local"
READ_DIST_MV_NAME = "sessions_hourly_mv_dist"
all_columns = ColumnSet(
[
("session_id", UUID()),
("distinct_id", UUID()),
("quantity", UInt(32)),
("seq", UInt(64)),
("org_id", UInt(64)),
("project_id", UInt(64)),
("retention_days", UInt(16)),
("duration", UInt(32)),
("status", UInt(8)),
("errors", UInt(16)),
("received", DateTime()),
("started", DateTime()),
("release", String()),
("environment", String()),
("user_agent", String()),
("os", String()),
IPv4,
IPv6,
Nested,
String,
UInt,
)
from snuba.clusters.storage_sets import StorageSetKey
from snuba.migrations import migration, operations, table_engines
from snuba.migrations.columns import Materialized
from snuba.migrations.columns import MigrationModifiers as Modifiers
UNKNOWN_SPAN_STATUS = 2
columns: List[Column[Modifiers]] = [
Column("project_id", UInt(64)),
Column("event_id", UUID()),
Column("trace_id", UUID()),
Column("span_id", UInt(64)),
Column("transaction_name", String(Modifiers(low_cardinality=True))),
Column(
"transaction_hash",
UInt(64, Modifiers(materialized="cityHash64(transaction_name)")),
),
Column("transaction_op", String(Modifiers(low_cardinality=True))),
Column("transaction_status",
UInt(8, Modifiers(default=str(UNKNOWN_SPAN_STATUS)))),
Column("start_ts", DateTime()),
Column("start_ms", UInt(16)),
Column("finish_ts", DateTime()),
Column("finish_ms", UInt(16)),
Column("duration", UInt(32)),
def __init__(self):
write_columns = ColumnSet([
('org_id', UInt(64)),
('project_id', UInt(64)),
('key_id', Nullable(UInt(64))),
('timestamp', DateTime()),
('outcome', UInt(8)),
('reason', LowCardinality(Nullable(String()))),
('event_id', Nullable(UUID())),
])
write_schema = MergeTreeSchema(
columns=write_columns,
# TODO: change to outcomes.raw_local when we add multi DB support
local_table_name=WRITE_LOCAL_TABLE_NAME,
dist_table_name=WRITE_DIST_TABLE_NAME,
order_by='(org_id, project_id, timestamp)',
partition_by='(toMonday(timestamp))',
settings={'index_granularity': 16384})
read_columns = ColumnSet([
('org_id', UInt(64)),
('project_id', UInt(64)),
('key_id', UInt(64)),
('timestamp', DateTime()),
('outcome', UInt(8)),
('reason', LowCardinality(String())),
('times_seen', UInt(64)),
])
read_schema = SummingMergeTreeSchema(
columns=read_columns,
local_table_name=READ_LOCAL_TABLE_NAME,
dist_table_name=READ_DIST_TABLE_NAME,
order_by='(org_id, project_id, key_id, outcome, reason, timestamp)',
partition_by='(toMonday(timestamp))',
settings={'index_granularity': 256})
materialized_view_columns = ColumnSet([
('org_id', UInt(64)),
('project_id', UInt(64)),
('key_id', UInt(64)),
('timestamp', DateTime()),
('outcome', UInt(8)),
('reason', String()),
('times_seen', UInt(64)),
])
# TODO: Find a better way to specify a query for a materialized view
# The problem right now is that we have a way to define our columns in a ColumnSet abstraction but the query
# doesn't use it.
query = """
SELECT
org_id,
project_id,
ifNull(key_id, 0) AS key_id,
toStartOfHour(timestamp) AS timestamp,
outcome,
ifNull(reason, 'none') AS reason,
count() AS times_seen
FROM %(source_table_name)s
GROUP BY org_id, project_id, key_id, timestamp, outcome, reason
"""
materialized_view = MaterializedViewSchema(
local_materialized_view_name='outcomes_mv_hourly_local',
dist_materialized_view_name='outcomes_mv_hourly_dist',
columns=materialized_view_columns,
query=query,
local_source_table_name=WRITE_LOCAL_TABLE_NAME,
local_destination_table_name=READ_LOCAL_TABLE_NAME,
dist_source_table_name=WRITE_DIST_TABLE_NAME,
dist_destination_table_name=READ_DIST_TABLE_NAME)
dataset_schemas = DatasetSchemas(
read_schema=read_schema,
write_schema=write_schema,
intermediary_schemas=[materialized_view])
super(OutcomesDataset, self).__init__(
dataset_schemas=dataset_schemas,
processor=OutcomesProcessor(),
default_topic="outcomes",
)
import pytest
from snuba.clickhouse.columns import UUID, UInt
from snuba.datasets.entities import EntityKey
from snuba.datasets.entities.entity_data_model import EntityColumnSet
from snuba.query import SelectedExpression
from snuba.query.composite import CompositeQuery
from snuba.query.conditions import ConditionFunctions, binary_condition
from snuba.query.data_source.simple import Entity
from snuba.query.expressions import Column, FunctionCall, Literal
from snuba.query.logical import Query
from snuba.utils.schemas import Column as EntityColumn
from snuba.web.query import ProjectsFinder
EVENTS_SCHEMA = EntityColumnSet([
EntityColumn("event_id", UUID()),
EntityColumn("project_id", UInt(32)),
EntityColumn("group_id", UInt(32)),
])
SIMPLE_QUERY = Query(
Entity(EntityKey.EVENTS, EVENTS_SCHEMA),
selected_columns=[
SelectedExpression(
"alias",
Column("_snuba_project", None, "project_id"),
)
],
array_join=None,
condition=binary_condition(
ConditionFunctions.IN,
Enum,
Float,
LowCardinality,
Nested,
Nullable,
String,
UInt,
UUID,
)
from snuba.clusters.storage_sets import StorageSetKey
from snuba.migrations import migration, operations, table_engines
status_type = Enum([("success", 0), ("error", 1), ("rate-limited", 2)])
columns = [
Column("request_id", UUID()),
Column("request_body", String()),
Column("referrer", LowCardinality(String())),
Column("dataset", LowCardinality(String())),
Column("projects", Array(UInt(64))),
Column("organization", Nullable(UInt(64))),
Column("timestamp", DateTime()),
Column("duration_ms", UInt(32)),
Column("status", status_type),
Column(
"clickhouse_queries",
Nested([
Column("sql", String()),
Column("status", status_type),
Column("trace_id", Nullable(UUID())),
Column("duration_ms", UInt(32)),
DateTime,
IPv4,
IPv6,
Nested,
String,
UInt,
)
from snuba.clusters.storage_sets import StorageSetKey
from snuba.datasets.storages.tags_hash_map import TAGS_HASH_MAP_COLUMN
from snuba.migrations import migration, operations, table_engines
from snuba.migrations.columns import MigrationModifiers as Modifiers
columns: Sequence[Column[Modifiers]] = [
Column("project_id", UInt(64)),
Column("timestamp", DateTime()),
Column("event_id", UUID(Modifiers(codecs=["NONE"]))),
Column("platform", String(Modifiers(low_cardinality=True))),
Column("environment", String(Modifiers(nullable=True,
low_cardinality=True))),
Column("release", String(Modifiers(nullable=True, low_cardinality=True))),
Column("dist", String(Modifiers(nullable=True, low_cardinality=True))),
Column("ip_address_v4", IPv4(Modifiers(nullable=True))),
Column("ip_address_v6", IPv6(Modifiers(nullable=True))),
Column("user", String(Modifiers(default="''"))),
Column("user_hash", UInt(64, Modifiers(materialized="cityHash64(user)"))),
Column("user_id", String(Modifiers(nullable=True))),
Column("user_name", String(Modifiers(nullable=True))),
Column("user_email", String(Modifiers(nullable=True))),
Column("sdk_name", String(Modifiers(nullable=True, low_cardinality=True))),
Column("sdk_version", String(Modifiers(nullable=True,
low_cardinality=True))),
from snuba.clusters.storage_sets import StorageSetKey
from snuba.datasets.replays_processor import ReplaysProcessor
from snuba.datasets.schemas.tables import WritableTableSchema
from snuba.datasets.storage import WritableTableStorage
from snuba.datasets.storages import StorageKey
from snuba.datasets.table_storage import build_kafka_stream_loader_from_settings
from snuba.query.processors.conditions_enforcer import ProjectIdEnforcer
from snuba.query.processors.table_rate_limit import TableRateLimit
from snuba.utils.schemas import Nested
from snuba.utils.streams.topics import Topic
LOCAL_TABLE_NAME = "replays_local"
DIST_TABLE_NAME = "replays_dist"
columns = ColumnSet([
("replay_id", UUID()),
("sequence_id", UInt(16)),
("timestamp", DateTime()),
(
"trace_ids",
Array(UUID()),
), # TODO: create bloom filter index / materialize column
("title", String(Modifiers(readonly=True))),
### common sentry event columns
("project_id", UInt(64)),
# release/environment info
("platform", String()),
("environment", String(Modifiers(nullable=True))),
("release", String(Modifiers(nullable=True))),
("dist", String(Modifiers(nullable=True))),
("ip_address_v4", IPv4(Modifiers(nullable=True))),
from typing import Sequence
from snuba.clickhouse.columns import (
UUID,
Column,
DateTime,
String,
UInt,
)
from snuba.clusters.storage_sets import StorageSetKey
from snuba.migrations import migration, operations, table_engines
from snuba.migrations.columns import MigrationModifiers as Modifiers
from .matview import aggregate_columns_v1, create_matview_v1
raw_columns: Sequence[Column[Modifiers]] = [
Column("session_id", UUID()),
Column("distinct_id", UUID()),
Column("seq", UInt(64)),
Column("org_id", UInt(64)),
Column("project_id", UInt(64)),
Column("retention_days", UInt(16)),
Column("duration", UInt(32)),
Column("status", UInt(8)),
Column("errors", UInt(16)),
Column("received", DateTime()),
Column("started", DateTime()),
Column("release", String(Modifiers(low_cardinality=True))),
Column("environment", String(Modifiers(low_cardinality=True))),
]
from snuba.datasets.storages.factory import get_writable_storage
from snuba.pipeline.simple_pipeline import SimplePipelineBuilder
from snuba.query.processors import QueryProcessor
from snuba.query.processors.object_id_rate_limiter import (
OrganizationRateLimiterProcessor,
ProjectRateLimiterProcessor,
ProjectReferrerRateLimiter,
ReferrerRateLimiterProcessor,
)
from snuba.query.processors.quota_processor import ResourceQuotaProcessor
from snuba.query.validation.validators import EntityRequiredColumnValidator
profile_columns = EntityColumnSet([
Column("organization_id", UInt(64)),
Column("project_id", UInt(64)),
Column("transaction_id", UUID()),
Column("profile_id", UUID()),
Column("received", DateTime()),
Column("profile", String()),
Column("android_api_level", UInt(32, Modifiers(nullable=True))),
Column("device_classification", String()),
Column("device_locale", String()),
Column("device_manufacturer", String()),
Column("device_model", String()),
Column("device_os_build_number", String(Modifiers(nullable=True))),
Column("device_os_name", String()),
Column("device_os_version", String()),
Column("duration_ns", UInt(64)),
Column("environment", String(Modifiers(nullable=True))),
Column("platform", String()),
Column("trace_id", UUID()),
from snuba.clickhouse.columns import UUID, Array, ColumnSet, DateTime, Float
from snuba.clickhouse.columns import SchemaModifiers as Modifiers
from snuba.clickhouse.columns import String, UInt
from snuba.clusters.storage_sets import StorageSetKey
from snuba.datasets.querylog_processor import QuerylogProcessor
from snuba.datasets.schemas.tables import WritableTableSchema
from snuba.datasets.storage import WritableTableStorage
from snuba.datasets.storages import StorageKey
from snuba.datasets.table_storage import build_kafka_stream_loader_from_settings
from snuba.utils.streams.topics import Topic
columns = ColumnSet([
("request_id", UUID()),
("request_body", String()),
("referrer", String()),
("dataset", String()),
("projects", Array(UInt(64))),
("organization", UInt(64, Modifiers(nullable=True))),
("timestamp", DateTime()),
("duration_ms", UInt(32)),
("status", String()),
# clickhouse_queries Nested columns.
# This is expanded into arrays instead of being expressed as a
# Nested column because, when adding new columns to a nested field
# we need to provide a default for the entire array (each new column
# is an array).
# The same schema cannot be achieved with the Nested construct (where
# we can only provide default for individual values), so, if we
# use the Nested construct, this schema cannot match the one generated
# by the migration framework (or by any ALTER statement).
("clickhouse_queries.sql", Array(String())),
from snuba.web.split import ColumnSplitQueryStrategy, TimeSplitQueryStrategy
required_columns = [
"event_id",
"primary_hash",
"project_id",
"group_id",
"timestamp",
"deleted",
"retention_days",
]
all_columns = ColumnSet([
("project_id", UInt(64)),
("timestamp", DateTime()),
("event_id", UUID()),
("platform", String()),
("environment", String(Modifiers(nullable=True))),
("release", String(Modifiers(nullable=True))),
("dist", String(Modifiers(nullable=True))),
("ip_address_v4", IPv4(Modifiers(nullable=True))),
("ip_address_v6", IPv6(Modifiers(nullable=True))),
("user", String()),
("user_hash", UInt(64, Modifiers(readonly=True))),
("user_id", String(Modifiers(nullable=True))),
("user_name", String(Modifiers(nullable=True))),
("user_email", String(Modifiers(nullable=True))),
("sdk_name", String(Modifiers(nullable=True))),
("sdk_version", String(Modifiers(nullable=True))),
("http_method", String(Modifiers(nullable=True))),
("http_referer", String(Modifiers(nullable=True))),
FixedString,
IPv4,
IPv6,
Nested,
String,
UInt,
)
from snuba.clusters.storage_sets import StorageSetKey
from snuba.migrations import migration, operations, table_engines
from snuba.migrations.columns import MigrationModifiers as Modifiers
columns: Sequence[Column[Modifiers]] = [
Column("org_id", UInt(64)),
Column("project_id", UInt(64)),
Column("timestamp", DateTime()),
Column("event_id", (UUID(Modifiers(codecs=["NONE"])))),
Column(
"event_hash",
UInt(
64,
Modifiers(materialized="cityHash64(toString(event_id))",
codecs=["NONE"]),
),
),
Column("platform", String(Modifiers(low_cardinality=True))),
Column("environment", String(Modifiers(nullable=True,
low_cardinality=True))),
Column("release", String(Modifiers(nullable=True, low_cardinality=True))),
Column("dist", String(Modifiers(nullable=True, low_cardinality=True))),
Column("ip_address_v4", IPv4(Modifiers(nullable=True))),
Column("ip_address_v6", IPv6(Modifiers(nullable=True))),
UInt,
WithDefault,
)
from snuba.clusters.storage_sets import StorageSetKey
from snuba.datasets.schemas.tables import WritableTableSchema
from snuba.datasets.spans_processor import UNKNOWN_SPAN_STATUS, SpansMessageProcessor
from snuba.datasets.storage import WritableTableStorage
from snuba.datasets.storages import StorageKey
from snuba.datasets.storages.tags_hash_map import TAGS_HASH_MAP_COLUMN
from snuba.datasets.table_storage import KafkaStreamLoader
from snuba.query.processors.prewhere import PrewhereProcessor
from snuba.web.split import TimeSplitQueryStrategy
columns = ColumnSet([
("project_id", UInt(64)),
("transaction_id", UUID()),
("trace_id", UUID()),
("transaction_span_id", UInt(64)),
("span_id", UInt(64)),
("parent_span_id", Nullable(UInt(64))),
("transaction_name", LowCardinality(String())),
("description", String()), # description in span
("op", LowCardinality(String())),
(
"status",
WithDefault(UInt(8), str(UNKNOWN_SPAN_STATUS)),
),
("start_ts", DateTime()),
("start_ns", UInt(32)),
("finish_ts", DateTime()),
("finish_ns", UInt(32)),
)
if "http_referer" not in current_schema:
ret.append(
f"ALTER TABLE {clickhouse_table} ADD COLUMN http_referer Nullable(String) AFTER http_method"
)
return ret
all_columns = ColumnSet(
[
("org_id", UInt(64)),
("project_id", UInt(64)),
("timestamp", DateTime()),
("event_id", WithCodecs(UUID(), ["NONE"])),
(
"event_hash",
WithCodecs(
Materialized(UInt(64), "cityHash64(toString(event_id))",), ["NONE"],
),
),
("platform", LowCardinality(String())),
("environment", LowCardinality(Nullable(String()))),
("release", LowCardinality(Nullable(String()))),
("dist", LowCardinality(Nullable(String()))),
("ip_address_v4", Nullable(IPv4())),
("ip_address_v6", Nullable(IPv6())),
("user", WithDefault(String(), "''")),
("user_hash", Materialized(UInt(64), "cityHash64(user)"),),
("user_id", Nullable(String())),
Column("project_id", UInt(64)),
Column("started", DateTime()),
Column("release", String(Modifiers(low_cardinality=True))),
Column("environment", String(Modifiers(low_cardinality=True))),
Column("user_agent", String(Modifiers(low_cardinality=True))),
Column("os", String(Modifiers(low_cardinality=True))),
# durations
Column(
"duration_quantiles",
AggregateFunction("quantilesIf(0.5, 0.9)",
[UInt(32), UInt(8)]),
),
Column("duration_avg", AggregateFunction("avgIf",
[UInt(32), UInt(8)])),
# sessions:
Column("sessions", AggregateFunction("countIf", [UUID(), UInt(8)])),
Column("sessions_preaggr", AggregateFunction("sumIf",
[UInt(32), UInt(8)])),
Column("sessions_crashed", AggregateFunction("countIf",
[UUID(), UInt(8)])),
Column("sessions_crashed_preaggr",
AggregateFunction("sumIf", [UInt(32), UInt(8)])),
Column("sessions_abnormal", AggregateFunction("countIf",
[UUID(), UInt(8)])),
Column("sessions_abnormal_preaggr",
AggregateFunction("sumIf", [UInt(32), UInt(8)])),
Column("sessions_errored", AggregateFunction("uniqIf",
[UUID(), UInt(8)])),
Column("sessions_errored_preaggr",
AggregateFunction("sumIf", [UInt(32), UInt(8)])),
# users:
from snuba.datasets.storages.errors_common import mandatory_conditions
from snuba.datasets.storages.events_bool_contexts import EventsBooleanContextsProcessor
from snuba.query.processors.arrayjoin_keyvalue_optimizer import (
ArrayJoinKeyValueOptimizer,
)
from snuba.query.processors.mapping_optimizer import MappingOptimizer
from snuba.query.processors.mapping_promoter import MappingColumnPromoter
from snuba.query.processors.prewhere import PrewhereProcessor
from snuba.query.processors.type_converters.uuid_column_processor import (
UUIDColumnProcessor,
)
from snuba.web.split import ColumnSplitQueryStrategy, TimeSplitQueryStrategy
columns = ColumnSet(
[
("event_id", UUID()),
("project_id", UInt(64)),
("type", String()),
("timestamp", DateTime()),
("platform", String()),
("environment", String(Modifiers(nullable=True))),
("release", String(Modifiers(nullable=True))),
("dist", String(Modifiers(nullable=True))),
("transaction_name", String()),
("message", String()),
("title", String()),
("user", String()),
("user_hash", UInt(64)),
("user_id", String(Modifiers(nullable=True))),
("user_name", String(Modifiers(nullable=True))),
("user_email", String(Modifiers(nullable=True))),
from snuba.clickhouse.columns import SchemaModifiers as Modifiers
from snuba.clickhouse.columns import String, UInt
from snuba.clusters.storage_sets import StorageSetKey
from snuba.datasets.schemas.tables import TableSchema
from snuba.datasets.storage import ReadableTableStorage
from snuba.datasets.storages import StorageKey
from snuba.datasets.storages.errors_common import mandatory_conditions
from snuba.datasets.storages.event_id_column_processor import EventIdColumnProcessor
from snuba.query.processors.arrayjoin_keyvalue_optimizer import (
ArrayJoinKeyValueOptimizer, )
from snuba.query.processors.mapping_optimizer import MappingOptimizer
from snuba.query.processors.prewhere import PrewhereProcessor
from snuba.web.split import TimeSplitQueryStrategy
columns = ColumnSet([
("event_id", UUID()),
("project_id", UInt(64)),
("type", String()),
("timestamp", DateTime()),
("platform", String()),
("environment", String(Modifiers(nullable=True))),
("release", String(Modifiers(nullable=True))),
("dist", String(Modifiers(nullable=True))),
("transaction_name", String()),
("message", String()),
("title", String()),
("user", String()),
("user_hash", UInt(64)),
("user_id", String(Modifiers(nullable=True))),
("user_name", String(Modifiers(nullable=True))),
("user_email", String(Modifiers(nullable=True))),
def __init__(self):
columns = ColumnSet([
('project_id', UInt(64)),
('event_id', UUID()),
('trace_id', UUID()),
('span_id', UInt(64)),
('transaction_name', String()),
('transaction_hash',
Materialized(
UInt(64),
'cityHash64(transaction_name)',
)),
('transaction_op', LowCardinality(String())),
('start_ts', DateTime()),
('start_ms', UInt(16)),
('finish_ts', DateTime()),
('finish_ms', UInt(16)),
('duration',
Materialized(
UInt(32),
'((finish_ts - start_ts) * 1000) + (finish_ms - start_ms)',
)),
('platform', LowCardinality(String())),
('environment', Nullable(String())),
('release', Nullable(String())),
('dist', Nullable(String())),
('ip_address_v4', Nullable(IPv4())),
('ip_address_v6', Nullable(IPv6())),
('user', WithDefault(
String(),
"''",
)),
('user_id', Nullable(String())),
('user_name', Nullable(String())),
('user_email', Nullable(String())),
('tags', Nested([
('key', String()),
('value', String()),
])),
('contexts', Nested([
('key', String()),
('value', String()),
])),
('partition', UInt(16)),
('offset', UInt(64)),
('retention_days', UInt(16)),
('deleted', UInt(8)),
])
schema = ReplacingMergeTreeSchema(
columns=columns,
local_table_name='transactions_local',
dist_table_name='transactions_dist',
order_by=
'(project_id, toStartOfDay(start_ts), transaction_hash, start_ts, start_ms, trace_id, span_id)',
partition_by='(retention_days, toMonday(start_ts))',
version_column='deleted',
sample_expr=None,
)
dataset_schemas = DatasetSchemas(
read_schema=schema,
write_schema=schema,
)
super().__init__(
dataset_schemas=dataset_schemas,
processor=TransactionsMessageProcessor(),
default_topic="events",
time_group_columns={
'bucketed_start': 'start_ts',
'bucketed_end': 'finish_ts',
},
)
ret.append(
f"ALTER TABLE {clickhouse_table} ADD COLUMN http_method LowCardinality(Nullable(String)) AFTER sdk_version"
)
if "http_referer" not in current_schema:
ret.append(
f"ALTER TABLE {clickhouse_table} ADD COLUMN http_referer Nullable(String) AFTER http_method"
)
return ret
columns = ColumnSet(
[
("project_id", UInt(64)),
("event_id", UUID()),
("trace_id", UUID()),
("span_id", UInt(64)),
("transaction_name", LowCardinality(String())),
("transaction_hash", Materialized(UInt(64), "cityHash64(transaction_name)",),),
("transaction_op", LowCardinality(String())),
("transaction_status", WithDefault(UInt(8), str(UNKNOWN_SPAN_STATUS))),
("start_ts", DateTime()),
("start_ms", UInt(16)),
("finish_ts", DateTime()),
("finish_ms", UInt(16)),
("duration", UInt(32)),
("platform", LowCardinality(String())),
("environment", LowCardinality(Nullable(String()))),
("release", LowCardinality(Nullable(String()))),
("dist", LowCardinality(Nullable(String()))),
def __init__(self) -> None:
self.__common_columns = ColumnSet(
[
("event_id", FixedString(32)),
("project_id", UInt(64)),
("type", Nullable(String())),
("timestamp", DateTime()),
("platform", Nullable(String())),
("environment", Nullable(String())),
("release", Nullable(String())),
("dist", Nullable(String())),
("user", Nullable(String())),
("transaction", Nullable(String())),
("message", Nullable(String())),
("title", Nullable(String())),
# User
("user_id", Nullable(String())),
("username", Nullable(String())),
("email", Nullable(String())),
("ip_address", Nullable(String())),
# SDK
("sdk_name", Nullable(String())),
("sdk_version", Nullable(String())),
# geo location context
("geo_country_code", Nullable(String())),
("geo_region", Nullable(String())),
("geo_city", Nullable(String())),
("http_method", Nullable(String())),
("http_referer", Nullable(String())),
# Other tags and context
("tags", Nested([("key", String()), ("value", String())])),
("contexts", Nested([("key", String()), ("value", String())])),
]
)
self.__events_columns = ColumnSet(
[
("group_id", Nullable(UInt(64))),
("primary_hash", Nullable(FixedString(32))),
# Promoted tags
("level", Nullable(String())),
("logger", Nullable(String())),
("server_name", Nullable(String())),
("site", Nullable(String())),
("url", Nullable(String())),
("search_message", Nullable(String())),
("location", Nullable(String())),
("culprit", Nullable(String())),
("received", Nullable(DateTime())),
("sdk_integrations", Nullable(Array(String()))),
("version", Nullable(String())),
# exception interface
(
"exception_stacks",
Nested(
[
("type", Nullable(String())),
("value", Nullable(String())),
("mechanism_type", Nullable(String())),
("mechanism_handled", Nullable(UInt(8))),
]
),
),
(
"exception_frames",
Nested(
[
("abs_path", Nullable(String())),
("filename", Nullable(String())),
("package", Nullable(String())),
("module", Nullable(String())),
("function", Nullable(String())),
("in_app", Nullable(UInt(8))),
("colno", Nullable(UInt(32))),
("lineno", Nullable(UInt(32))),
("stack_level", UInt(16)),
]
),
),
("modules", Nested([("name", String()), ("version", String())])),
]
)
self.__transactions_columns = ColumnSet(
[
("trace_id", Nullable(UUID())),
("span_id", Nullable(UInt(64))),
("transaction_hash", Nullable(UInt(64))),
("transaction_op", Nullable(String())),
("transaction_status", Nullable(UInt(8))),
("duration", Nullable(UInt(32))),
(
"measurements",
Nested([("key", LowCardinality(String())), ("value", Float(64))]),
),
]
)
events_storage = get_storage(StorageKey.EVENTS)
events_ro_storage = get_storage(StorageKey.EVENTS_RO)
transactions_storage = get_storage(StorageKey.TRANSACTIONS)
self.__time_group_columns: Mapping[str, str] = {}
self.__time_parse_columns = ("timestamp",)
super().__init__(
storages=[events_storage, transactions_storage],
query_plan_builder=SelectedStorageQueryPlanBuilder(
selector=DiscoverQueryStorageSelector(
events_table=events_storage,
events_ro_table=events_ro_storage,
abstract_events_columns=self.__events_columns,
transactions_table=transactions_storage,
abstract_transactions_columns=self.__transactions_columns,
),
),
abstract_column_set=(
self.__common_columns
+ self.__events_columns
+ self.__transactions_columns
),
writable_storage=None,
)
