:hedgehog: PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP.
Python ab-testingai-analyticsanalyticscdpdata-warehouseexperimentsfeature-flagsjavascriptproduct-analyticspythonreactsession-replaysurveystypescriptweb-analytics
from datetime import datetime
from typing import TYPE_CHECKING
from posthog.hogql import ast
from posthog.hogql.parser import parse_select
from posthog.hogql.query import execute_hogql_query
if TYPE_CHECKING:
from posthog.models import Team
# One page of spans can carry hundreds of distinct ids. The lookup is one query either way,
# but the IN list grows with it, so cap how many ids one request can ask about.
MAX_IDS_PER_LOOKUP = 200
def _run(query: ast.SelectQuery | ast.SelectSetQuery, team: "Team", query_type: str) -> dict[str, int]:
response = execute_hogql_query(query=query, team=team, query_type=query_type)
return {str(key): int(exceptions) for key, exceptions in response.results or []}
def _window_placeholders(date_from: datetime, date_to: datetime) -> dict[str, ast.Expr]:
return {
"date_from": ast.Constant(value=date_from),
"date_to": ast.Constant(value=date_to),
# Explicit, because HogQL caps a select without a LIMIT at 100 rows, which is below the
# ids one request may ask about. A row cut there would read as a clean id.
"limit": ast.Constant(value=MAX_IDS_PER_LOOKUP),
}
def count_trace_exceptions(
*, team: "Team", trace_ids: list[str], date_from: datetime, date_to: datetime
) -> dict[str, int]:
"""Exceptions Error Tracking linked to an issue, per trace, inside the window.
Keys come back lowercased. A trace with no such exceptions is absent from the result rather
than present with a zero.
"""
if not trace_ids:
return {}
query = parse_select(
"""
SELECT lower(properties.$trace_id) AS trace_id, count() AS exceptions
FROM events
WHERE event = '$exception'
AND isNotNull(properties.$exception_issue_id)
AND timestamp >= {date_from}
AND timestamp <= {date_to}
AND lower(properties.$trace_id) IN {trace_ids}
GROUP BY trace_id
LIMIT {limit}
""",
placeholders={
# Spans read their ids back as uppercase hex, while the SDKs write them lowercase, so
# both sides are lowercased here. Compare the two raw and every count comes back zero.
"trace_ids": ast.Constant(value=[trace_id.lower() for trace_id in trace_ids]),
**_window_placeholders(date_from, date_to),
},
)
return _run(query, team, "TracingTraceErrorCountsQuery")
def count_span_exceptions(
*, team: "Team", span_ids: list[str], trace_ids: list[str], date_from: datetime, date_to: datetime
) -> dict[str, int]:
"""Exceptions Error Tracking linked to an issue, per span, inside the window.
A span id is only unique within its trace, so the traces the spans belong to bound the match
as well. Keys come back lowercased.
"""
if not span_ids or not trace_ids:
return {}
query = parse_select(
"""
SELECT lower(properties.$span_id) AS span_id, count() AS exceptions
FROM events
WHERE event = '$exception'
AND isNotNull(properties.$exception_issue_id)
AND timestamp >= {date_from}
AND timestamp <= {date_to}
AND lower(properties.$trace_id) IN {trace_ids}
AND lower(properties.$span_id) IN {span_ids}
GROUP BY span_id
LIMIT {limit}
""",
placeholders={
"span_ids": ast.Constant(value=[span_id.lower() for span_id in span_ids]),
"trace_ids": ast.Constant(value=[trace_id.lower() for trace_id in trace_ids]),
**_window_placeholders(date_from, date_to),
},
)
return _run(query, team, "TracingSpanErrorCountsQuery")
def count_session_exceptions(
*, team: "Team", session_ids: list[str], date_from: datetime, date_to: datetime
) -> dict[str, int]:
"""Exceptions Error Tracking linked to an issue, per session, inside the window.
A session with no such exceptions is absent from the result rather than present with a zero.
"""
if not session_ids:
return {}
query = parse_select(
"""
SELECT properties.$session_id AS session_id, count() AS exceptions
FROM events
WHERE event = '$exception'
AND isNotNull(properties.$exception_issue_id)
AND timestamp >= {date_from}
AND timestamp <= {date_to}
AND properties.$session_id IN {session_ids}
GROUP BY session_id
LIMIT {limit}
""",
placeholders={
"session_ids": ast.Constant(value=session_ids),
**_window_placeholders(date_from, date_to),
},
)
return _run(query, team, "TracingSessionErrorCountsQuery")