chore: import upstream snapshot with attribution

surfsense_backend/scripts/docker/entrypoint.e2e.sh

@@ -0,0 +1,53 @@
+#!/bin/bash
+# =============================================================================
+# E2E entrypoint for the multi-stage Dockerfile's `e2e` target.
+#
+# Dispatches on SERVICE_ROLE to the test-only entrypoints under tests/e2e/.
+# Those scripts apply sys.modules hijacks and LLM/embedding patches BEFORE
+# importing production app code (see tests/e2e/run_backend.py for rationale).
+#
+# Production never sees this file: tests/ is excluded from the production
+# stage, and the production stage uses scripts/docker/entrypoint.sh.
+# =============================================================================
+set -euo pipefail
+
+SERVICE_ROLE="${SERVICE_ROLE:-api}"
+echo "[e2e-entrypoint] starting role=${SERVICE_ROLE}"
+
+wait_for_db() {
+    # Block until the database is reachable. We don't loop forever — Compose
+    # depends_on/healthchecks already gate on db readiness, this is just
+    # belt-and-suspenders so a slow first connection doesn't race migrations.
+    for i in {1..60}; do
+        echo "[e2e-entrypoint] db check attempt ${i}/60"
+        if python -c "from app.db import engine; import asyncio; asyncio.run(engine.dispose())"; then
+            echo "[e2e-entrypoint] db reachable after ${i} attempts"
+            return 0
+        fi
+        sleep 1
+    done
+    echo "[e2e-entrypoint] ERROR: db not reachable after 60s" >&2
+    return 1
+}
+
+case "${SERVICE_ROLE}" in
+    api)
+        wait_for_db
+        echo "[e2e-entrypoint] running alembic upgrade head"
+        alembic upgrade head
+        # `exec` so SIGTERM from `docker stop` reaches Python directly,
+        # without a shell wrapper interposing.
+        exec python tests/e2e/run_backend.py
+        ;;
+    worker)
+        # Worker doesn't run migrations — the api role does that exactly once.
+        # We still wait for db so Celery's broker connection check doesn't
+        # race against an unready Postgres on cold start.
+        wait_for_db
+        exec python tests/e2e/run_celery.py
+        ;;
+    *)
+        echo "[e2e-entrypoint] ERROR: unknown SERVICE_ROLE='${SERVICE_ROLE}' (expected: api | worker)" >&2
+        exit 1
+        ;;
+esac

surfsense_backend/scripts/docker/entrypoint.sh

@@ -0,0 +1,160 @@
+#!/bin/bash
+set -e
+
+# ─────────────────────────────────────────────────────────────
+# SERVICE_ROLE controls which process(es) this container runs.
+#
+#   migrate – Run `alembic upgrade head`, verify zero_publication,
+#             then exit 0. Used by the dedicated `migrations` service
+#             in docker-compose.yml so downstream services can gate
+#             on `condition: service_completed_successfully`.
+#   api     – FastAPI backend only (does NOT run migrations)
+#   worker  – Celery worker only
+#   beat    – Celery beat scheduler only
+#   all     – migrations + api + worker + beat in one container
+#             (legacy / dev default; fails fast on migration error)
+#
+# Set SERVICE_ROLE as an environment variable in Coolify for
+# each service deployment.
+# ─────────────────────────────────────────────────────────────
+SERVICE_ROLE="${SERVICE_ROLE:-all}"
+echo "Starting SurfSense with SERVICE_ROLE=${SERVICE_ROLE}"
+
+# ── Autoscale defaults (override via env) ────────────────────
+#   CELERY_MAX_WORKERS  – max concurrent worker processes
+#   CELERY_MIN_WORKERS  – min workers kept warm
+#   CELERY_QUEUES       – comma-separated queues to consume
+#                         (empty = all queues for backward compat)
+CELERY_MAX_WORKERS="${CELERY_MAX_WORKERS:-10}"
+CELERY_MIN_WORKERS="${CELERY_MIN_WORKERS:-2}"
+CELERY_MAX_TASKS_PER_CHILD="${CELERY_MAX_TASKS_PER_CHILD:-50}"
+CELERY_QUEUES="${CELERY_QUEUES:-}"
+
+# ── Graceful shutdown ────────────────────────────────────────
+PIDS=()
+
+cleanup() {
+    echo "Shutting down services..."
+    for pid in "${PIDS[@]}"; do
+        kill -TERM "$pid" 2>/dev/null || true
+    done
+    for pid in "${PIDS[@]}"; do
+        wait "$pid" 2>/dev/null || true
+    done
+    exit 0
+}
+
+trap cleanup SIGTERM SIGINT
+
+# ── Database migrations (only for migrate / all) ─────────────
+# Fail-fast contract:
+#   - alembic upgrade head must succeed within ${MIGRATION_TIMEOUT:-900}s
+#   - zero_publication must match the canonical app.zero_publication shape
+# Either failure exits non-zero so the dedicated `migrations` compose
+# service exits non-zero, halting the rest of the stack instead of
+# silently producing a drifted Zero publication.
+run_migrations() {
+    echo "Running database migrations..."
+    for i in {1..30}; do
+        if python -c "from app.db import engine; import asyncio; asyncio.run(engine.dispose())" 2>/dev/null; then
+            echo "Database is ready."
+            break
+        fi
+        echo "Waiting for database... ($i/30)"
+        sleep 1
+    done
+
+    local timeout_secs="${MIGRATION_TIMEOUT:-900}"
+    echo "Running alembic upgrade head (timeout=${timeout_secs}s)..."
+    if ! timeout "${timeout_secs}" alembic upgrade head; then
+        echo "ERROR: alembic upgrade head failed (or exceeded ${timeout_secs}s timeout)." >&2
+        echo "Refusing to start. Inspect the error above and re-run." >&2
+        exit 1
+    fi
+    echo "Migrations completed successfully."
+
+    echo "Verifying zero_publication matches the canonical shape..."
+    if ! python -m app.zero_publication --verify; then
+        echo "ERROR: zero_publication does not match the canonical shape." >&2
+        echo "Inspect alembic state with:" >&2
+        echo "  docker compose exec db psql -U \"\$DB_USER\" -d \"\$DB_NAME\" -c 'SELECT * FROM alembic_version;'" >&2
+        exit 1
+    fi
+}
+
+# ── Service starters ─────────────────────────────────────────
+start_api() {
+    echo "Starting FastAPI Backend..."
+    python main.py &
+    PIDS+=($!)
+    echo "  FastAPI PID=${PIDS[-1]}"
+}
+
+start_worker() {
+    QUEUE_ARGS=""
+    if [ -n "${CELERY_QUEUES}" ]; then
+        QUEUE_ARGS="--queues=${CELERY_QUEUES}"
+    else
+        # When no queues specified, consume from the default, connectors, and
+        # gateway maintenance queues. Without --queues, Celery only consumes
+        # from the default queue, leaving connector/gateway maintenance tasks stuck.
+        DEFAULT_Q="${CELERY_TASK_DEFAULT_QUEUE:-surfsense}"
+        QUEUE_ARGS="--queues=${DEFAULT_Q},${DEFAULT_Q}.connectors,${DEFAULT_Q}.gateway"
+    fi
+
+    echo "Starting Celery Worker (autoscale=${CELERY_MAX_WORKERS},${CELERY_MIN_WORKERS}, max-tasks-per-child=${CELERY_MAX_TASKS_PER_CHILD}, queues=${CELERY_QUEUES:-all})..."
+    celery -A app.celery_app worker \
+        --loglevel=info \
+        --autoscale="${CELERY_MAX_WORKERS},${CELERY_MIN_WORKERS}" \
+        --max-tasks-per-child="${CELERY_MAX_TASKS_PER_CHILD}" \
+        --prefetch-multiplier=1 \
+        -Ofair \
+        ${QUEUE_ARGS} &
+    PIDS+=($!)
+    echo "  Celery Worker PID=${PIDS[-1]}"
+}
+
+start_beat() {
+    echo "Starting Celery Beat..."
+    celery -A app.celery_app beat --loglevel=info &
+    PIDS+=($!)
+    echo "  Celery Beat PID=${PIDS[-1]}"
+}
+
+# ── Main: run based on role ──────────────────────────────────
+case "${SERVICE_ROLE}" in
+    migrate)
+        run_migrations
+        echo "Migrations complete; exiting cleanly."
+        exit 0
+        ;;
+    api)
+        start_api
+        ;;
+    worker)
+        start_worker
+        ;;
+    beat)
+        start_beat
+        ;;
+    all)
+        run_migrations
+        start_api
+        sleep 5
+        start_worker
+        sleep 3
+        start_beat
+        ;;
+    *)
+        echo "ERROR: Unknown SERVICE_ROLE '${SERVICE_ROLE}'. Use: migrate, api, worker, beat, or all"
+        exit 1
+        ;;
+esac
+
+echo "All requested services started. PIDs: ${PIDS[*]}"
+
+# Wait for any process to exit
+wait -n
+
+# If we get here, one process exited unexpectedly
+exit $?