fix(sidebar): prime idx_messages_session before CLI-session scan (#3887) (#3888)

* fix(sidebar): prime idx_messages_session before CLI-session scan (#3887)

The sidebar's CLI-session scan (read_importable_agent_session_rows) orders
candidate sessions by a correlated MAX(timestamp) subquery over messages,
which collapses to a full per-session table scan when the agent's standard
idx_messages_session ON messages(session_id, timestamp) index is absent.
A normally-migrated hermes-agent state.db has it; a db that lost its
migrations (older hermes-agent, or a hand-rebuilt/reimported db) does not,
and /api/sessions then stalls for seconds on every refresh (the 5s cache
TTL never settles -> repeating 'Slow WebUI request still running' warnings).

Prime the index with CREATE INDEX IF NOT EXISTS before the scan: a no-op
when it already exists, a ~20ms self-heal otherwise (measured 13.3s ->
0.009s on a no-index 8k-session db). Best-effort: degrades silently on a
read-only db, a locked db, or a minimal schema without a timestamp column.

Closes #3887

* test(#3887): skip read-only-db test under root (greptile P2)

Root bypasses POSIX permission bits, so chmod 0444 doesn't make the file
read-only for root — the prime would succeed and the test would validate the
wrong path, giving false confidence on root-run CI. Skip under root; the
production handler's except sqlite3.Error: pass covers the read-only/locked/
corrupted/older-schema cases regardless.

---------

Co-authored-by: nesquena-hermes <[email protected]>
This commit is contained in:
nesquena-hermes
2026-06-09 12:35:13 -07:00
committed by GitHub
parent 35098632ea
commit 5d4d7fe9fd
3 changed files with 206 additions and 0 deletions

View File

@@ -3,6 +3,10 @@
## [Unreleased]
### Fixed
- **Sidebar no longer stalls for seconds on a state.db that is missing the agent's message index.** The CLI-session scan orders candidate sessions by a correlated `MAX(timestamp)` subquery over `messages`, which collapses to a full per-session table scan when `idx_messages_session ON messages(session_id, timestamp)` is absent — stalling `/api/sessions` for several seconds on every refresh (the 5s cache TTL never settles, so repeated "Slow WebUI request still running" warnings appear). A normally-migrated hermes-agent state.db has this index, but a db that lost its migrations (older hermes-agent, or a hand-rebuilt/reimported db) does not. WebUI now primes the index with `CREATE INDEX IF NOT EXISTS` before the scan — a no-op when it already exists, and a ~20ms self-heal otherwise (measured 13.3s → 0.009s on a no-index 8k-session db). Best-effort: degrades silently on a read-only db, a locked db, or a minimal schema without a `timestamp` column. (#3887)
## [v0.51.341] — 2026-06-09 — Release LE (stale thinking-dot placeholder fix)
### Fixed

View File

@@ -462,6 +462,28 @@ def read_importable_agent_session_rows(
use_messages_join = messages_has_session_id
count_col = 'id' if 'id' in message_cols else 'session_id'
# Defensive index prime (#3887). The candidate-ordering query below sorts
# sessions by a correlated ``MAX(mx.timestamp)`` subquery over ``messages``.
# That is fast only when the agent's standard
# ``idx_messages_session ON messages(session_id, timestamp)`` index exists.
# A normally-migrated hermes-agent state.db has it, but a db that lost its
# migrations (older hermes-agent, or a hand-rebuilt/reimported db) does
# not — and the subquery then degrades to a full ``messages`` scan per
# candidate session, stalling ``/api/sessions`` for seconds on every
# refresh (the 5s-TTL cache never settles). Priming the index is a no-op
# (~free) when it already exists, and self-heals an affected db in
# milliseconds. Best-effort: degrade silently on a read-only db or any
# error so the listing never fails because of the prime.
if messages_has_session_id and messages_has_timestamp:
try:
cur.execute(
"CREATE INDEX IF NOT EXISTS idx_messages_session "
"ON messages(session_id, timestamp)"
)
conn.commit()
except sqlite3.Error:
pass # read-only db / locked / older schema — degrade gracefully
if use_messages_join:
actual_count_expr = f"COUNT(m.{count_col})"
if 'role' in message_cols:

View File

@@ -0,0 +1,180 @@
"""Regression tests for #3887 — defensive index prime for the sidebar scan.
The sidebar's CLI-session scan (``read_importable_agent_session_rows``) orders
candidate sessions by a correlated ``MAX(mx.timestamp)`` subquery over the
``messages`` table. That is fast only when the agent's standard
``idx_messages_session ON messages(session_id, timestamp)`` index exists. A
state.db that lost its migrations (older hermes-agent, or a hand-rebuilt /
reimported db) has no such index and the scan degrades to a full ``messages``
scan per candidate session — stalling ``/api/sessions`` for seconds on every
refresh.
These tests assert the intent (the index is primed when missing so the listing
self-heals) and the cross-cell isolation (the prime is a no-op when the index
already exists, is skipped when the schema lacks the columns, and degrades
silently on a read-only db without ever failing the listing).
"""
import os
import sqlite3
import stat
import pytest
import api.agent_sessions as agent_sessions
def _full_schema_db(path):
"""A state.db with the columns the scan reads, but NO messages index."""
conn = sqlite3.connect(str(path))
cur = conn.cursor()
cur.execute(
"""CREATE TABLE sessions(
id TEXT PRIMARY KEY, title TEXT, model TEXT, message_count INTEGER,
started_at REAL, source TEXT, session_source TEXT,
parent_session_id TEXT, ended_at REAL, end_reason TEXT,
user_id TEXT, chat_id TEXT, chat_type TEXT, thread_id TEXT,
session_key TEXT, origin_chat_id TEXT, origin_user_id TEXT,
platform TEXT)"""
)
cur.execute(
"""CREATE TABLE messages(
id INTEGER PRIMARY KEY AUTOINCREMENT, session_id TEXT, role TEXT,
content TEXT, timestamp REAL)"""
)
for i in range(3):
sid = f"sess{i}"
cur.execute(
"INSERT INTO sessions(id, title, model, message_count, started_at, "
"source, session_source) VALUES (?,?,?,?,?,?,?)",
(sid, f"Session {i}", "m", 2, 1000.0 + i, "cli", "cli"),
)
cur.execute(
"INSERT INTO messages(session_id, role, content, timestamp) "
"VALUES (?,?,?,?)",
(sid, "user", "hi", 1001.0 + i),
)
cur.execute(
"INSERT INTO messages(session_id, role, content, timestamp) "
"VALUES (?,?,?,?)",
(sid, "assistant", "yo", 1002.0 + i),
)
conn.commit()
conn.close()
def _messages_indexes(path):
conn = sqlite3.connect(str(path))
try:
rows = conn.execute(
"SELECT name FROM sqlite_master "
"WHERE type='index' AND tbl_name='messages'"
).fetchall()
finally:
conn.close()
return {r[0] for r in rows}
def test_prime_creates_missing_index(tmp_path):
"""A db missing idx_messages_session gets it primed on the first scan."""
db = tmp_path / "state.db"
_full_schema_db(db)
assert "idx_messages_session" not in _messages_indexes(db)
rows = agent_sessions.read_importable_agent_session_rows(
db, limit=20, exclude_sources=None
)
# Listing still returns the sessions ...
assert {r["id"] for r in rows} == {"sess0", "sess1", "sess2"}
# ... and the index now exists on (session_id, timestamp).
assert "idx_messages_session" in _messages_indexes(db)
conn = sqlite3.connect(str(db))
try:
sql = conn.execute(
"SELECT sql FROM sqlite_master WHERE name='idx_messages_session'"
).fetchone()[0]
finally:
conn.close()
assert "session_id" in sql and "timestamp" in sql
def test_prime_is_noop_when_index_exists(tmp_path):
"""When the agent already created the index, the prime must not duplicate
or error — the existing index is reused untouched."""
db = tmp_path / "state.db"
_full_schema_db(db)
conn = sqlite3.connect(str(db))
conn.execute(
"CREATE INDEX idx_messages_session ON messages(session_id, timestamp)"
)
conn.commit()
conn.close()
before = _messages_indexes(db)
rows = agent_sessions.read_importable_agent_session_rows(
db, limit=20, exclude_sources=None
)
assert {r["id"] for r in rows} == {"sess0", "sess1", "sess2"}
# Index set is unchanged — no duplicate, no error.
assert _messages_indexes(db) == before
assert "idx_messages_session" in _messages_indexes(db)
def test_prime_skipped_without_timestamp_column(tmp_path):
"""A minimal messages schema (no timestamp column) must not attempt the
index and must not crash the listing."""
db = tmp_path / "state.db"
conn = sqlite3.connect(str(db))
cur = conn.cursor()
cur.execute(
"""CREATE TABLE sessions(
id TEXT PRIMARY KEY, title TEXT, model TEXT, message_count INTEGER,
started_at REAL, source TEXT)"""
)
cur.execute(
"CREATE TABLE messages(id INTEGER PRIMARY KEY, session_id TEXT, "
"role TEXT, content TEXT)"
)
cur.execute(
"INSERT INTO sessions(id, title, model, message_count, started_at, "
"source) VALUES ('s1','T','m',1,1000.0,'cli')"
)
cur.execute(
"INSERT INTO messages(id, session_id, role, content) "
"VALUES (1,'s1','user','hi')"
)
conn.commit()
conn.close()
rows = agent_sessions.read_importable_agent_session_rows(
db, limit=20, exclude_sources=None
)
# Listing degrades gracefully via the denormalized counts (still surfaces).
assert {r["id"] for r in rows} == {"s1"}
# The timestamp-less schema must NOT have an index primed on it.
assert "idx_messages_session" not in _messages_indexes(db)
def test_prime_degrades_on_readonly_db(tmp_path):
"""A read-only db (can't create the index) must not raise — the listing
still returns, just without the perf benefit."""
# Root bypasses POSIX permission bits, so chmod 0444 doesn't make the file
# read-only for root and the prime would succeed — validating the wrong path
# and giving false confidence on root-run CI (greptile). Skip under root; the
# production handler's `except sqlite3.Error: pass` covers read-only/locked/
# corrupted/older-schema regardless, and the other cases pin that contract.
if hasattr(os, "getuid") and os.getuid() == 0:
pytest.skip("chmod-based read-only test is a no-op under root")
db = tmp_path / "state.db"
_full_schema_db(db)
os.chmod(db, stat.S_IREAD)
try:
rows = agent_sessions.read_importable_agent_session_rows(
db, limit=20, exclude_sources=None
)
assert {r["id"] for r in rows} == {"sess0", "sess1", "sess2"}
finally:
# Restore write so tmp_path cleanup can remove it.
os.chmod(db, stat.S_IWRITE | stat.S_IREAD)