Keep the engine's state in SQLite, not Postgres

One process owns this database — the image has run a single uvicorn
worker for that reason since the four-engines bug — so a file beside the
flows is the honest shape for it, and it is what lets `fluksio serve`
need no infrastructure at all. Live values, node execution and the work
queue never came here anyway; what does is a rollup a minute at a time,
a row per cascade and the run history, and WAL keeps the readers going
while that one writer works.

DATA_DIR is now the one setting that moves everything an installation
keeps; the rest derive from it and the images still spell theirs out.
The schema is prepared in-process at startup, so the prestart service is
gone, and the ten Postgres-only revisions collapse into one portable
baseline.

Three things only worked because psycopg was casting for us: a token's
subject arriving as a string where the column is a UUID, `greatest`, and
`date_bin`. The timestamps needed a column type of their own — SQLite
stores no offset, and a naive datetime read back either raises against an
aware `now` or serialises as local time.

Postgres stays in the stack only for Umami, behind the analytics profile.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

backend/tests/__init__.py

@@ -1,14 +1,32 @@
 """Redirect the suite at its own database.
 
-`app.core.db` builds the engine at import time and several modules bind that
-object, so the name has to be in the environment before anything imports the
-settings. Overriding it here — the first module pytest imports for the
+`fluksio.core.db` builds the engine at import time and several modules bind
+that object, so the URL has to be in the environment before anything imports
+the settings. Overriding it here — the first module pytest imports for the
 package — keeps a test run from touching the development data.
+
+A file rather than `:memory:`: the suite starts the engine's own threads (the
+run service, the metrics collector, the portal connector) against this engine,
+and an in-memory database is per connection unless every one of them shares a
+single connection, which those threads would then contend for.
 """
 
 import os
+import tempfile
 
-os.environ["POSTGRES_DB"] = "app_test"
+_DB_DIR = tempfile.mkdtemp(prefix="fluksio-test-")
+os.environ.setdefault("DATABASE_URL", f"sqlite:///{_DB_DIR}/app_test.db")
+# Nothing outside this directory: the suite must not pick up a checkout's .env
+# and write into the development data.
+os.environ.setdefault("FLUKSIO_ENV_FILE", os.path.join(_DB_DIR, "env"))
+os.environ.setdefault("DATA_DIR", _DB_DIR)
+# Named here rather than taken from a developer's .env, so the suite is the
+# same run everywhere and a checkout's own superuser is never involved.
+os.environ.setdefault("FIRST_SUPERUSER", "admin@example.com")
+os.environ.setdefault("FIRST_SUPERUSER_PASSWORD", "testpassword")
+# What `emails_enabled` needs; the tests that send mail patch SMTP_HOST on top
+# of it. Nothing is delivered — the sender is faked where it matters.
+os.environ.setdefault("EMAILS_FROM_EMAIL", "noreply@example.com")
 # The MCP session manager can only be entered once per instance, and the suite
 # builds a TestClient — and so a lifespan — per test module. Tests that want the
 # endpoint mount it themselves.