Skip to content

Self-hosted monitoring for homelabs: host and container metrics and uptime checks from all your machines in one dashboard, with an optional AI agent that diagnoses incidents and suggests fixes.

feat(db): Alembic migrations with a baseline that brings existing databases up to date

karloks2005bb8343692119
diff --git a/TESTS.md b/TESTS.md
index 9c53b69..c187661 100644
--- a/TESTS.md
+++ b/TESTS.md
@@ -164,6 +164,7 @@ The chart helper tests for the frontend are separate, in `src/frontend/tests/`.
| [`test_health_checks.py`](#test_health_checkspy) | 4 | The assistant's fixed health checks and the critical-only webhook. |
| [`test_history.py`](#test_historypy) | 36 | The history API behind long-range charts: which level of detail answers each range, point limits, gaps, and agreement with the raw samples. |
| [`test_indexes.py`](#test_indexespy) | 9 | The database indexes on the metrics tables: only the ones queries use, none duplicated. |
+| [`test_migrations.py`](#test_migrationspy) | 5 | Database migrations: new and pre-migration databases both end at the latest revision with the models' schema. |
| [`test_ingest_validation.py`](#test_ingest_validationpy) | 178 | What collectors report is checked and bounded: numbers, names, filesystems, spacing of check-ins. |
| [`test_input_limits.py`](#test_input_limitspy) | 38 | Size and shape limits on request bodies, the audit-log endpoint, and safe defaults. |
| [`test_network_checks.py`](#test_network_checkspy) | 94 | The host's real network totals, listening-port inventory, optional internet check, firewall fingerprints, and the validation of all of it. |
@@ -682,6 +683,19 @@ The indexes on the metrics tables. The live migration is checked by hand against
| `test_deleting_a_server_and_purging_old_rows_can_both_use_an_index` *(4 inputs)* | Every metrics table keeps an index starting with `server_id` (deleting a server cascades through it) and one on `time` alone (the hourly purge). |
| `test_a_primary_key_is_never_indexed_a_second_time` *(4 inputs)* | No table gets a redundant index on its own `id`. |
+### test_migrations.py
+
+The Alembic migrations in `src/backend/migrations`. Every test runs on SQLite, and also on PostgreSQL when
+`MIGRATION_TEST_POSTGRES_URL` points at an empty database (its schema is wiped first).
+
+| Test | What it checks |
+|---|---|
+| `test_the_migrations_live_next_to_the_backend` | The revisions are where the backend looks for them. |
+| `test_a_new_database_ends_at_the_latest_revision_with_the_models_schema` | An empty database ends at the latest revision, and Alembic finds no difference between it and the models. |
+| `test_upgrading_twice_changes_nothing` | Running the upgrade again at the latest revision changes no table. |
+| `test_a_database_from_before_migrations_keeps_its_data_and_ends_at_the_latest_revision` | A database made the old way (`create_all`) keeps its rows and ends at the latest revision with the models' schema. |
+| `test_an_old_database_missing_tables_and_columns_is_brought_up_to_date` | An install from an older version, without the availability tables and some `servers` columns, gets them. |
+
### test_ingest_validation.py
Nothing a collector reports is trusted: numbers, names and filesystems are checked, check-ins are spaced, and one bad collector can't break pages for everyone.
diff --git a/alembic.ini b/alembic.ini
new file mode 100644
index 0000000..50a6375
--- /dev/null
+++ b/alembic.ini
@@ -0,0 +1,33 @@
+# Database migrations (Alembic). The backend runs them itself at start (src/backend/migrate.py); this file is for
+# writing new ones: `alembic revision --autogenerate -m "what changed"` against a database at the latest revision.
+# The database comes from DATABASE_URL, as for the backend. See docs/development.md.
+[alembic]
+script_location = %(here)s/src/backend/migrations
+file_template = %%(rev)s_%%(slug)s
+prepend_sys_path = .
+
+[loggers]
+keys = root,alembic
+
+[handlers]
+keys = console
+
+[formatters]
+keys = generic
+
+[logger_root]
+level = WARNING
+handlers = console
+
+[logger_alembic]
+level = INFO
+handlers =
+qualname = alembic
+
+[handler_console]
+class = StreamHandler
+args = (sys.stderr,)
+formatter = generic
+
+[formatter_generic]
+format = %(levelname)s [%(name)s] %(message)s
diff --git a/compose.yml b/compose.yml
index 8f022b2..b230e3c 100644
--- a/compose.yml
+++ b/compose.yml
@@ -11,7 +11,6 @@ services:
POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:?Set POSTGRES_PASSWORD in .env}
volumes:
- postgres_data:/var/lib/postgresql/data
- - ./docker/postgres/init:/docker-entrypoint-initdb.d
healthcheck:
test: ["CMD-SHELL", "pg_isready -U $${POSTGRES_USER:-infradeck} -d $${POSTGRES_DB:-infradeck}"]
interval: 10s
diff --git a/docker-compose.server.yml b/docker-compose.server.yml
index 50bef8f..082c90b 100644
--- a/docker-compose.server.yml
+++ b/docker-compose.server.yml
@@ -16,7 +16,6 @@ services:
POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:?Set POSTGRES_PASSWORD in .env}
volumes:
- postgres_data:/var/lib/postgresql/data
- - ./docker/postgres/init:/docker-entrypoint-initdb.d
healthcheck:
test: ["CMD-SHELL", "pg_isready -U $${POSTGRES_USER:-infradeck} -d $${POSTGRES_DB:-infradeck}"]
interval: 10s
diff --git a/docker/postgres/init/01_create_users.sql b/docker/postgres/init/01_create_users.sql
deleted file mode 100644
index 9b1a67c..0000000
--- a/docker/postgres/init/01_create_users.sql
+++ /dev/null
@@ -1,13 +0,0 @@
-CREATE TABLE IF NOT EXISTS users (
- id SERIAL PRIMARY KEY,
- email VARCHAR(255) UNIQUE NOT NULL,
- password_hash TEXT NOT NULL,
- name VARCHAR(255) NOT NULL,
- role VARCHAR(50) NOT NULL DEFAULT 'user',
- is_active BOOLEAN NOT NULL DEFAULT TRUE,
- created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
- updated_at TIMESTAMPTZ NOT NULL DEFAULT NOW()
-);
-
-CREATE INDEX IF NOT EXISTS idx_users_email ON users (email);
-CREATE INDEX IF NOT EXISTS idx_users_role ON users (role);
diff --git a/docs/containers.md b/docs/containers.md
index bee8352..d06dec7 100644
--- a/docs/containers.md
+++ b/docs/containers.md
@@ -140,8 +140,8 @@ while it was down are lost (the collector does not queue them), so the history h
The full list of tables is in [Database schema](database.md).
**How it works.** The official `postgres:16-alpine` image with its data in the `postgres_data` volume.
-On the very first start, [`docker/postgres/init/`](../docker/postgres/init) creates the users table; the
-backend creates and upgrades everything else itself at startup. The healthcheck (`pg_isready`) holds the
+The backend creates every table and upgrades the schema itself at startup (Alembic migrations in
+[`src/backend/migrations/`](../src/backend/migrations)). The healthcheck (`pg_isready`) holds the
backend back until the database accepts connections. The container is on the internal `data` network
only and publishes no port.
diff --git a/docs/database.md b/docs/database.md
index 152c676..d7c558c 100644
--- a/docs/database.md
+++ b/docs/database.md
@@ -2,9 +2,8 @@
# Database schema
-Tables are created on backend startup (`Base.metadata.create_all`), and new
-columns are added with `ALTER TABLE ... IF NOT EXISTS` in the startup hook in
-`src/backend/main.py`. Each check-in stores one row in `host_metrics` (which also
+The backend creates and upgrades the schema itself at startup with Alembic migrations
+(see [Migrations](#migrations)). Each check-in stores one row in `host_metrics` (which also
carries the boot ID, OOM counter, kernel flags, clock state and last shutdown), one
per filesystem in `mount_metrics` and one per container in `container_metrics` (which
also records the exit code, out-of-memory flag, finish time and image ID), all with the
@@ -12,8 +11,8 @@ same timestamp. `system_events` holds reboots, OOM kills, kernel messages, conta
changes (appeared, removed, image changed) and network changes (ports opened or closed,
firewall changes); `listener_states` and `firewall_states` hold what was last seen, `join_tokens` the hashed single-use tokens, `recovery_codes` the hashed two-factor
recovery codes (the encrypted two-factor secret lives on the `users` row), and per-mount disk
-thresholds live in `app_settings`. `docker/postgres/init/01_create_users.sql` initialises
-the `users` table on a fresh volume.
+thresholds live in `app_settings`. Every table is created and changed by the migrations in
+`src/backend/migrations/` (see Migrations below).
**The agent's spending.** `agent_runs` has one row per call to the language model: what it was for (`chat`, later
`diagnosis`), whether it worked, the model, input and output tokens, the estimated cost in USD, how long it took, and
@@ -39,9 +38,29 @@ purge never deletes samples that have not been summarised yet. The Statistics pa
not purged yet; a retention setting for each level is planned.
The metrics tables carry only the indexes their queries use: `time` (for the hourly purge) and
-`(server_id, time)`, or `(server_id, container_id, time)` for containers. Older databases are
-migrated at start: the new indexes are created and the redundant single-column ones are dropped,
-which can be repeated safely. Creating an index briefly blocks writes to that table (a few
-seconds for a table of millions of rows); on a very large install, create the indexes yourself
-first with `CREATE INDEX CONCURRENTLY IF NOT EXISTS <name> ON <table> (<columns>)` using the names
-in `src/backend/models.py`, and the start-up migration will skip them.
+`(server_id, time)`, or `(server_id, container_id, time)` for containers. A database from before
+migrations gets them from the baseline migration, which also drops the redundant single-column
+ones. Creating an index briefly blocks writes to that table (a few seconds for a table of millions
+of rows); on a very large install, create the indexes yourself first with
+`CREATE INDEX CONCURRENTLY IF NOT EXISTS <name> ON <table> (<columns>)` using the names in
+`src/backend/models.py`, and the migration will skip them.
+
+## Migrations
+
+The schema is versioned with [Alembic](https://alembic.sqlalchemy.org/). The revisions live in
+`src/backend/migrations/versions/`, and the database records the one it is at in the `alembic_version`
+table. At every start the backend runs the revisions the database doesn't have yet, in order, before it
+serves anything (`src/backend/migrate.py`):
+
+- On PostgreSQL an advisory lock makes a second process wait instead of migrating at the same time.
+- Each revision runs in its own transaction. If one fails, the database stays at the revision before it and
+ the backend refuses to start; the log says which revision failed.
+- The log has one line per revision applied (`Database migration applied: 0002 (...)`).
+
+The first revision, `0001_baseline`, is the schema as it was when migrations were introduced. A new database
+gets every table from it. A database from before migrations (made by the backend's old start-up code) keeps
+its tables and data: the baseline creates only the tables it lacks and brings the others up to date, so
+upgrading needs no manual step.
+
+How to add a revision is in [Development](development.md#database-migrations).
+
diff --git a/docs/development.md b/docs/development.md
index d1a2762..84f0c12 100644
--- a/docs/development.md
+++ b/docs/development.md
@@ -49,6 +49,22 @@ helpers have their own tests that need only Node 20 or newer: `cd src/frontend &
[TESTS.md](../TESTS.md) for how the tests work, what they don't cover, and what every
test is for.
+## Database migrations
+
+Change the models in `src/backend/models.py`, then generate a revision against a database that is at the
+latest revision (the development stack's PostgreSQL, or any empty database the backend has started on once):
+
+```bash
+DATABASE_URL=postgresql+psycopg://infradeck:<password>@127.0.0.1:5432/infradeck alembic revision --autogenerate -m "add servers.location"
+```
+
+Read the generated file in `src/backend/migrations/versions/` before committing it: Alembic can't tell a
+renamed column from a dropped and added one, and data changes (backfills) have to be written by hand. Name
+revisions with the next number (`--rev-id 0002`). The backend applies new revisions at its next start;
+`alembic upgrade head` and `alembic downgrade -1` do it by hand. `tests/history/test_migrations.py` fails when
+the models and the revisions describe different schemas; set `MIGRATION_TEST_POSTGRES_URL` to an empty
+PostgreSQL database to run it there as well as on SQLite.
+
## Project layout
```text
@@ -61,6 +77,7 @@ SECURITY.md what each component can touch, and its limits
TESTS.md how to run the tests and what each one checks
requirements-dev.txt app requirements plus pytest, for running the tests
pytest.ini test settings (lets plain `pytest` find the `src` package)
+alembic.ini settings for the `alembic` command (writing new migrations)
.github/workflows/
images.yml builds, signs and publishes the collector images on a version tag
docker/
@@ -73,12 +90,13 @@ docker/
nginx.strict.conf server stack: public listener, collector endpoints refused
nginx.collector.conf server stack: collector gate (VPN and internal listeners)
Caddyfile optional HTTPS front door (dev compose, --profile https)
- postgres/init/ SQL run on a fresh PostgreSQL volume
src/
backend/ API: auth, monitoring, enrollment, rate limits, models
alerts.py alert engine: one incident per rule, auto-resolve
rules.py alert rules and default thresholds
rollups.py summarises check-ins into hourly/daily/weekly/monthly rollups
+ migrate.py runs the database migrations at start
+ migrations/ Alembic revisions (versions/); 0001 is the baseline
history.py range-aware history behind the Statistics page
stats.py fleet statistics: incidents over time, availability, busy-hours heatmap
totp.py one-time codes and recovery codes (standard library only)
diff --git a/requirements.txt b/requirements.txt
index a09676d..2bd2d00 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,3 +1,4 @@
+alembic
bcrypt
cryptography
email-validator
diff --git a/src/backend/main.py b/src/backend/main.py
index 1762bfd..6c9f652 100644
--- a/src/backend/main.py
+++ b/src/backend/main.py
@@ -22,7 +22,7 @@ from fastapi import Path as PathParam
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import JSONResponse, PlainTextResponse
from pydantic import BaseModel, EmailStr, Field, field_validator
-from sqlalchemy import case, func, inspect, text
+from sqlalchemy import case, func
from .alerts import CONTAINER_RULE_PREFIX, CRITICAL, SERVER_OFFLINE_RULE, SEVERITIES, active_rule_keys, raise_alert, resolve_alerts, sync_alerts
import contextvars
@@ -34,9 +34,10 @@ from .auth import (
register_user, revoke_other_sessions, revoke_user_session, validate_password, verify_credentials, verify_password, verify_second_factor,
)
from .db import SessionLocal, check_database, get_db
-from .models import ApiToken, Invite, NotificationChannel, AppSetting, AuditLog, Base, ChatMessage, Check, ContainerMetric, ContainerRollup, Conversation, EnrollmentKey, FirewallState, HostMetric, HostRollup, IncidentAlert, JoinToken, ListenerState, MountMetric, MountRollup, PendingCollector, RecoveryCode, Server, SystemEvent, User
+from .models import ApiToken, Invite, NotificationChannel, AppSetting, AuditLog, ChatMessage, Check, ContainerMetric, ContainerRollup, Conversation, EnrollmentKey, FirewallState, HostMetric, HostRollup, IncidentAlert, JoinToken, ListenerState, MountMetric, MountRollup, PendingCollector, RecoveryCode, Server, SystemEvent, User
from .db import engine
+from .migrate import upgrade_database
from .config import (
AGENT_MODEL,
APP_VERSION,
@@ -330,83 +331,7 @@ def registration_warning(db) -> str | None:
@app.on_event("startup")
def startup_event():
- Base.metadata.create_all(bind=engine)
- with engine.begin() as connection:
- if engine.dialect.name == "postgresql":
- connection.execute(text("ALTER TABLE audit_logs ADD COLUMN IF NOT EXISTS task_title VARCHAR(255)"))
- connection.execute(text("ALTER TABLE audit_logs ADD COLUMN IF NOT EXISTS steps TEXT"))
- connection.execute(text("ALTER TABLE audit_logs ADD COLUMN IF NOT EXISTS conversation_id VARCHAR(255)"))
- connection.execute(text("ALTER TABLE audit_logs ADD COLUMN IF NOT EXISTS event_type VARCHAR(50)"))
- for column, column_type in (
- ("memory_limit_mb", "DOUBLE PRECISION"), ("net_rx_mb", "DOUBLE PRECISION"), ("net_tx_mb", "DOUBLE PRECISION"),
- ("block_read_mb", "DOUBLE PRECISION"), ("block_write_mb", "DOUBLE PRECISION"), ("pids", "INTEGER"),
- ("started_at", "VARCHAR(64)"), ("ports", "VARCHAR(512)"),
- ):
- connection.execute(text(f"ALTER TABLE container_metrics ADD COLUMN IF NOT EXISTS {column} {column_type}"))
- for column, column_type in (
- ("boot_id", "VARCHAR(64)"), ("oom_kill_total", "INTEGER"), ("kernel_tainted", "INTEGER"),
- ("ntp_synced", "BOOLEAN"), ("reboot_required", "BOOLEAN"), ("last_shutdown", "VARCHAR(16)"),
- ("docker_status", "VARCHAR(16)"), ("docker_reason", "VARCHAR(16)"), ("docker_latency_ms", "INTEGER"),
- ("internet_status", "VARCHAR(16)"), ("internet_latency_ms", "INTEGER"), ("collect_duration_ms", "INTEGER"),
- ):
- connection.execute(text(f"ALTER TABLE host_metrics ADD COLUMN IF NOT EXISTS {column} {column_type}"))
- for column, column_type in (("exit_code", "INTEGER"), ("oom_killed", "BOOLEAN"), ("finished_at", "TIMESTAMP WITH TIME ZONE"), ("image_id", "VARCHAR(80)"), ("project", "VARCHAR(128)"), ("volumes", "VARCHAR(512)"), ("swap_mb", "DOUBLE PRECISION")):
- connection.execute(text(f"ALTER TABLE container_metrics ADD COLUMN IF NOT EXISTS {column} {column_type}"))
- connection.execute(text("ALTER TABLE system_events ADD COLUMN IF NOT EXISTS event_key VARCHAR(128)"))
- connection.execute(text("ALTER TABLE pending_collectors ADD COLUMN IF NOT EXISTS join_token_id INTEGER REFERENCES join_tokens(id) ON DELETE SET NULL"))
- connection.execute(text("ALTER TABLE pending_collectors ADD COLUMN IF NOT EXISTS details TEXT"))
- for column, column_type in (("avg_swap_mb", "DOUBLE PRECISION"), ("n_swap_mb", "INTEGER NOT NULL DEFAULT 0"), ("max_swap_mb", "DOUBLE PRECISION")):
- connection.execute(text(f"ALTER TABLE container_rollups ADD COLUMN IF NOT EXISTS {column} {column_type}"))
- connection.execute(text("CREATE INDEX IF NOT EXISTS ix_system_events_event_key ON system_events (event_key)"))
- # Composite indexes that match how the metrics tables are read, then the single-column ones they make
- # redundant (and the pointless indexes on each table's own primary key). Safe to repeat. Creating an index
- # briefly blocks writes to that table; on a very large table create them first by hand with CONCURRENTLY.
- for table, name, columns in (
- ("host_metrics", "ix_host_metrics_server_time", "server_id, time"),
- ("mount_metrics", "ix_mount_metrics_server_time", "server_id, time"),
- ("system_events", "ix_system_events_server_time", "server_id, time"),
- ("container_metrics", "ix_container_metrics_server_container_time", "server_id, container_id, time"),
- ):
- connection.execute(text(f"CREATE INDEX IF NOT EXISTS {name} ON {table} ({columns})"))
- for redundant in (
- "ix_host_metrics_id", "ix_host_metrics_server_id", "ix_mount_metrics_id", "ix_mount_metrics_server_id",
- "ix_system_events_id", "ix_system_events_server_id",
- "ix_container_metrics_id", "ix_container_metrics_server_id", "ix_container_metrics_container_id",
- ):
- connection.execute(text(f"DROP INDEX IF EXISTS {redundant}"))
- for column, column_type in (("totp_secret_enc", "TEXT"), ("totp_enabled", "BOOLEAN NOT NULL DEFAULT FALSE"), ("totp_last_step", "INTEGER"), ("password_changed_at", "TIMESTAMP WITH TIME ZONE")):
- connection.execute(text(f"ALTER TABLE users ADD COLUMN IF NOT EXISTS {column} {column_type}"))
- connection.execute(text("ALTER TABLE incident_alerts ADD COLUMN IF NOT EXISTS rule_key VARCHAR(255)"))
- connection.execute(text("ALTER TABLE incident_alerts ADD COLUMN IF NOT EXISTS resolved_at TIMESTAMP WITH TIME ZONE"))
- for column, column_type in (("resolution", "VARCHAR(255)"), ("acknowledged_at", "TIMESTAMP WITH TIME ZONE"), ("acknowledged_by", "INTEGER"), ("muted_until", "TIMESTAMP WITH TIME ZONE"),
- ("notified_at", "TIMESTAMP WITH TIME ZONE"), ("resolve_notified_at", "TIMESTAMP WITH TIME ZONE"), ("reopen_after", "TIMESTAMP WITH TIME ZONE")):
- connection.execute(text(f"ALTER TABLE incident_alerts ADD COLUMN IF NOT EXISTS {column} {column_type}"))
- for column, column_type in (("username", "VARCHAR(64)"), ("avatar_color", "VARCHAR(16)"), ("preferences", "TEXT"), ("last_seen_at", "TIMESTAMP WITH TIME ZONE"), ("agent_chat_enabled", "BOOLEAN NOT NULL DEFAULT FALSE")):
- connection.execute(text(f"ALTER TABLE users ADD COLUMN IF NOT EXISTS {column} {column_type}"))
- connection.execute(text("CREATE INDEX IF NOT EXISTS ix_incident_alerts_rule_key ON incident_alerts (rule_key)"))
- # Alerts from before the alert engine: offline alerts keep going under their rule, the rest were
- # duplicated on every check-in with no way to resolve, so close them; the engine re-raises live ones.
- connection.execute(text("UPDATE incident_alerts SET rule_key = :rule WHERE rule_key IS NULL AND description = :description"), {"rule": SERVER_OFFLINE_RULE, "description": OFFLINE_ALERT_DESCRIPTION})
- connection.execute(text("UPDATE incident_alerts SET status = 'resolved', resolved_at = NOW() WHERE rule_key IS NULL AND status = 'active'"))
- # Disk and inode alerts became per mount; the old keys were for the root filesystem.
- connection.execute(text("UPDATE incident_alerts SET rule_key = rule_key || ':/' WHERE rule_key IN ('host:disk', 'host:inodes')"))
- if "servers" in inspect(engine).get_table_names():
- columns = {column["name"] for column in inspect(engine).get_columns("servers")}
- if "collector_token_hash" not in columns:
- with engine.begin() as connection:
- connection.execute(text("ALTER TABLE servers ADD COLUMN collector_token_hash VARCHAR(64)"))
- if "machine_id" not in columns:
- with engine.begin() as connection:
- connection.execute(text("ALTER TABLE servers ADD COLUMN machine_id VARCHAR(64)"))
- if "is_dashboard_host" not in columns:
- with engine.begin() as connection:
- connection.execute(text("ALTER TABLE servers ADD COLUMN is_dashboard_host BOOLEAN NOT NULL DEFAULT FALSE"))
- if "system_info" not in columns:
- with engine.begin() as connection:
- connection.execute(text("ALTER TABLE servers ADD COLUMN system_info TEXT"))
- if "name_locked" not in columns:
- with engine.begin() as connection:
- connection.execute(text("ALTER TABLE servers ADD COLUMN name_locked BOOLEAN NOT NULL DEFAULT FALSE"))
+ upgrade_database(engine) # src/backend/migrations; a failing revision stops the start here
db = SessionLocal()
try:
ensure_admin_exists(db)
diff --git a/src/backend/migrate.py b/src/backend/migrate.py
new file mode 100644
index 0000000..dc8c3af
--- /dev/null
+++ b/src/backend/migrate.py
@@ -0,0 +1,40 @@
+"""Bringing the database to the latest schema at start (Alembic revisions in src/backend/migrations/versions).
+
+Runs before the backend serves anything. On PostgreSQL an advisory lock makes a second process (another backend, or
+the agent worker) wait instead of migrating at the same time. Each revision runs in its own transaction, so a
+revision that fails leaves the database at the one before it, and the backend refuses to start rather than run
+on a half-changed schema.
+"""
+from __future__ import annotations
+
+from pathlib import Path
+
+from alembic import command
+from alembic.config import Config
+from sqlalchemy import text
+
+# Any fixed number; it only has to be the same for every process that migrates this database.
+ADVISORY_LOCK_ID = 7_120_351
+MIGRATIONS_DIR = Path(__file__).resolve().parent / "migrations"
+
+
+def alembic_config(connection) -> Config:
+ config = Config()
+ config.set_main_option("script_location", str(MIGRATIONS_DIR))
+ config.attributes["connection"] = connection
+ return config
+
+
+def upgrade_database(engine) -> None:
+ with engine.connect() as connection:
+ postgres = connection.dialect.name == "postgresql"
+ if postgres:
+ connection.execute(text("SELECT pg_advisory_lock(:id)"), {"id": ADVISORY_LOCK_ID})
+ connection.commit()
+ try:
+ command.upgrade(alembic_config(connection), "head")
+ connection.commit()
+ finally:
+ if postgres:
+ connection.execute(text("SELECT pg_advisory_unlock(:id)"), {"id": ADVISORY_LOCK_ID})
+ connection.commit()
diff --git a/src/backend/migrations/env.py b/src/backend/migrations/env.py
new file mode 100644
index 0000000..0a8c940
--- /dev/null
+++ b/src/backend/migrations/env.py
@@ -0,0 +1,28 @@
+"""Alembic's entry point. The backend passes its own connection in (src/backend/migrate.py); the `alembic` command
+line connects to DATABASE_URL instead."""
+from alembic import context
+from sqlalchemy import create_engine
+
+from src.backend.config import DATABASE_URL
+from src.backend.models import Base
+
+
+def _applied(ctx, step, heads, run_args) -> None:
+ print(f"Database migration {'applied' if step.is_upgrade else 'reverted'}: {step.up_revision_id} ({step.up_revision.doc.splitlines()[0]})", flush=True)
+
+
+def _run(connection) -> None:
+ # SQLite can't alter most of a table in place; batch mode copies the table instead. One transaction per
+ # revision, so a failing one leaves the database at the revision before it.
+ context.configure(connection=connection, target_metadata=Base.metadata, render_as_batch=connection.dialect.name == "sqlite", compare_type=True,
+ transaction_per_migration=True, on_version_apply=_applied)
+ with context.begin_transaction():
+ context.run_migrations()
+
+
+connection = context.config.attributes.get("connection")
+if connection is not None:
+ _run(connection)
+else:
+ with create_engine(DATABASE_URL).connect() as own:
+ _run(own)
diff --git a/src/backend/migrations/script.py.mako b/src/backend/migrations/script.py.mako
new file mode 100644
index 0000000..efe908a
--- /dev/null
+++ b/src/backend/migrations/script.py.mako
@@ -0,0 +1,20 @@
+"""${message}
+
+Revision ${up_revision}, after ${down_revision | comma,n}.
+"""
+from alembic import op
+import sqlalchemy as sa
+${imports if imports else ""}
+
+revision = ${repr(up_revision)}
+down_revision = ${repr(down_revision)}
+branch_labels = ${repr(branch_labels)}
+depends_on = ${repr(depends_on)}
+
+
+def upgrade() -> None:
+ ${upgrades if upgrades else "pass"}
+
+
+def downgrade() -> None:
+ ${downgrades if downgrades else "pass"}
diff --git a/src/backend/migrations/versions/0001_baseline.py b/src/backend/migrations/versions/0001_baseline.py
new file mode 100644
index 0000000..21c411d
--- /dev/null
+++ b/src/backend/migrations/versions/0001_baseline.py
@@ -0,0 +1,696 @@
+"""The schema as it was when migrations were introduced.
+
+A new database gets every table here. A database from before Alembic (created by ``create_all`` and the hand-written
+statements that used to run at start) keeps its tables: only the tables it lacks are created, and the old statements
+below bring the rest up to this schema. Both end at the same schema, which tests/history/test_migrations.py checks.
+"""
+from alembic import op
+import sqlalchemy as sa
+
+
+revision = '0001'
+down_revision = None
+branch_labels = None
+depends_on = None
+
+_existing: set[str] = set()
+
+
+def _table(name, *elements, **kwargs) -> None:
+ if name not in _existing:
+ op.create_table(name, *elements, **kwargs)
+
+
+def _index(name, table, columns, **kwargs) -> None:
+ if table not in _existing:
+ op.create_index(name, table, columns, **kwargs)
+
+
+def upgrade() -> None:
+ bind = op.get_bind()
+ _existing.clear()
+ _existing.update(sa.inspect(bind).get_table_names())
+ _create_tables()
+ if _existing:
+ _bring_up_to_date(bind)
+
+
+def _create_tables() -> None:
+ _table('app_settings',
+ sa.Column('key', sa.String(length=100), nullable=False),
+ sa.Column('value', sa.Text(), nullable=False),
+ sa.Column('updated_at', sa.DateTime(timezone=True), nullable=False),
+ sa.PrimaryKeyConstraint('key')
+ )
+ _table('audit_logs',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('user_id', sa.Integer(), nullable=True),
+ sa.Column('level', sa.String(length=50), nullable=False),
+ sa.Column('message', sa.Text(), nullable=False),
+ sa.Column('source', sa.String(length=100), nullable=False),
+ sa.Column('task_title', sa.String(length=255), nullable=True),
+ sa.Column('steps', sa.Text(), nullable=True),
+ sa.Column('conversation_id', sa.String(length=255), nullable=True),
+ sa.Column('event_type', sa.String(length=50), nullable=True),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.PrimaryKeyConstraint('id')
+ )
+ _index(op.f('ix_audit_logs_conversation_id'), 'audit_logs', ['conversation_id'], unique=False)
+ _index(op.f('ix_audit_logs_id'), 'audit_logs', ['id'], unique=False)
+ _index(op.f('ix_audit_logs_user_id'), 'audit_logs', ['user_id'], unique=False)
+ _table('chat_messages',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('conversation_id', sa.String(length=64), nullable=False),
+ sa.Column('role', sa.String(length=20), nullable=False),
+ sa.Column('content', sa.Text(), nullable=False),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.PrimaryKeyConstraint('id')
+ )
+ _index(op.f('ix_chat_messages_conversation_id'), 'chat_messages', ['conversation_id'], unique=False)
+ _index(op.f('ix_chat_messages_id'), 'chat_messages', ['id'], unique=False)
+ _table('checks',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('name', sa.String(length=120), nullable=False),
+ sa.Column('type', sa.String(length=16), nullable=False),
+ sa.Column('target', sa.String(length=2048), nullable=False),
+ sa.Column('options', sa.Text(), nullable=False),
+ sa.Column('tags', sa.String(length=255), nullable=False),
+ sa.Column('server_ids', sa.Text(), nullable=False),
+ sa.Column('rule', sa.String(length=8), nullable=False),
+ sa.Column('interval_seconds', sa.Integer(), nullable=False),
+ sa.Column('retries', sa.Integer(), nullable=False),
+ sa.Column('retry_interval_seconds', sa.Integer(), nullable=False),
+ sa.Column('timeout_seconds', sa.Integer(), nullable=False),
+ sa.Column('severity', sa.String(length=16), nullable=False),
+ sa.Column('paused', sa.Boolean(), nullable=False),
+ sa.Column('state', sa.String(length=16), nullable=False),
+ sa.Column('state_since', sa.DateTime(timezone=True), nullable=True),
+ sa.Column('last_result_at', sa.DateTime(timezone=True), nullable=True),
+ sa.Column('rolled_until', sa.DateTime(timezone=True), nullable=True),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.PrimaryKeyConstraint('id')
+ )
+ _table('conversations',
+ sa.Column('id', sa.String(length=64), nullable=False),
+ sa.Column('user_id', sa.Integer(), nullable=False),
+ sa.Column('title', sa.String(length=255), nullable=False),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('updated_at', sa.DateTime(timezone=True), nullable=False),
+ sa.PrimaryKeyConstraint('id')
+ )
+ _index(op.f('ix_conversations_user_id'), 'conversations', ['user_id'], unique=False)
+ _table('users',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('email', sa.String(length=255), nullable=False),
+ sa.Column('name', sa.String(length=255), nullable=False),
+ sa.Column('password_hash', sa.Text(), nullable=False),
+ sa.Column('role', sa.String(length=50), nullable=False),
+ sa.Column('is_active', sa.Boolean(), nullable=False),
+ sa.Column('totp_secret_enc', sa.Text(), nullable=True),
+ sa.Column('totp_enabled', sa.Boolean(), nullable=False),
+ sa.Column('totp_last_step', sa.Integer(), nullable=True),
+ sa.Column('password_changed_at', sa.DateTime(timezone=True), nullable=True),
+ sa.Column('username', sa.String(length=64), nullable=True),
+ sa.Column('avatar_color', sa.String(length=16), nullable=True),
+ sa.Column('preferences', sa.Text(), nullable=True),
+ sa.Column('agent_chat_enabled', sa.Boolean(), nullable=False),
+ sa.Column('last_seen_at', sa.DateTime(timezone=True), nullable=True),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('updated_at', sa.DateTime(timezone=True), nullable=False),
+ sa.PrimaryKeyConstraint('id')
+ )
+ _index(op.f('ix_users_email'), 'users', ['email'], unique=True)
+ _index(op.f('ix_users_id'), 'users', ['id'], unique=False)
+ _table('agent_runs',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('kind', sa.String(length=16), nullable=False),
+ sa.Column('status', sa.String(length=16), nullable=False),
+ sa.Column('model', sa.String(length=100), nullable=False),
+ sa.Column('input_tokens', sa.Integer(), nullable=False),
+ sa.Column('output_tokens', sa.Integer(), nullable=False),
+ sa.Column('cost_usd', sa.Float(), nullable=True),
+ sa.Column('duration_ms', sa.Integer(), nullable=False),
+ sa.Column('user_id', sa.Integer(), nullable=True),
+ sa.Column('conversation_id', sa.String(length=255), nullable=True),
+ sa.Column('incident_id', sa.Integer(), nullable=True),
+ sa.Column('error', sa.String(length=300), nullable=True),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.ForeignKeyConstraint(['user_id'], ['users.id'], ondelete='SET NULL'),
+ sa.PrimaryKeyConstraint('id')
+ )
+ _index('ix_agent_runs_created', 'agent_runs', ['created_at'], unique=False)
+ _index(op.f('ix_agent_runs_user_id'), 'agent_runs', ['user_id'], unique=False)
+ _table('api_tokens',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('user_id', sa.Integer(), nullable=False),
+ sa.Column('name', sa.String(length=64), nullable=False),
+ sa.Column('scope', sa.String(length=16), nullable=False),
+ sa.Column('token_hash', sa.String(length=64), nullable=False),
+ sa.Column('hint', sa.String(length=8), nullable=False),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('last_used_at', sa.DateTime(timezone=True), nullable=True),
+ sa.Column('expires_at', sa.DateTime(timezone=True), nullable=True),
+ sa.ForeignKeyConstraint(['user_id'], ['users.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('token_hash')
+ )
+ _index(op.f('ix_api_tokens_id'), 'api_tokens', ['id'], unique=False)
+ _index(op.f('ix_api_tokens_user_id'), 'api_tokens', ['user_id'], unique=False)
+ _table('check_results',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('check_id', sa.Integer(), nullable=False),
+ sa.Column('server_id', sa.Integer(), nullable=False),
+ sa.Column('time', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('ok', sa.Boolean(), nullable=False),
+ sa.Column('response_ms', sa.Integer(), nullable=True),
+ sa.Column('code', sa.Integer(), nullable=True),
+ sa.Column('detail', sa.String(length=200), nullable=True),
+ sa.Column('cert_days', sa.Integer(), nullable=True),
+ sa.ForeignKeyConstraint(['check_id'], ['checks.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id')
+ )
+ _index('ix_check_results_check_time', 'check_results', ['check_id', 'time'], unique=False)
+ _table('check_rollups',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('check_id', sa.Integer(), nullable=False),
+ sa.Column('hour', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('slots', sa.Integer(), nullable=False),
+ sa.Column('down_slots', sa.Integer(), nullable=False),
+ sa.Column('avg_ms', sa.Float(), nullable=True),
+ sa.Column('max_ms', sa.Integer(), nullable=True),
+ sa.ForeignKeyConstraint(['check_id'], ['checks.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('check_id', 'hour')
+ )
+ _index(op.f('ix_check_rollups_check_id'), 'check_rollups', ['check_id'], unique=False)
+ _table('enrollment_keys',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('owner_id', sa.Integer(), nullable=False),
+ sa.Column('key_hash', sa.String(length=64), nullable=False),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.ForeignKeyConstraint(['owner_id'], ['users.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('key_hash'),
+ sa.UniqueConstraint('owner_id')
+ )
+ _index(op.f('ix_enrollment_keys_id'), 'enrollment_keys', ['id'], unique=False)
+ _table('invites',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('email', sa.String(length=255), nullable=False),
+ sa.Column('role', sa.String(length=50), nullable=False),
+ sa.Column('token_hash', sa.String(length=64), nullable=False),
+ sa.Column('created_by', sa.Integer(), nullable=False),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('expires_at', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('used_at', sa.DateTime(timezone=True), nullable=True),
+ sa.ForeignKeyConstraint(['created_by'], ['users.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('token_hash')
+ )
+ _index(op.f('ix_invites_id'), 'invites', ['id'], unique=False)
+ _table('join_tokens',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('token_hash', sa.String(length=64), nullable=False),
+ sa.Column('created_by', sa.Integer(), nullable=False),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('expires_at', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('used_at', sa.DateTime(timezone=True), nullable=True),
+ sa.ForeignKeyConstraint(['created_by'], ['users.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('token_hash')
+ )
+ _index(op.f('ix_join_tokens_id'), 'join_tokens', ['id'], unique=False)
+ _table('notification_channels',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('user_id', sa.Integer(), nullable=False),
+ sa.Column('kind', sa.String(length=16), nullable=False),
+ sa.Column('target_enc', sa.Text(), nullable=False),
+ sa.Column('display', sa.String(length=120), nullable=False),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('last_status', sa.String(length=64), nullable=True),
+ sa.Column('last_ok', sa.Boolean(), nullable=True),
+ sa.Column('last_checked_at', sa.DateTime(timezone=True), nullable=True),
+ sa.ForeignKeyConstraint(['user_id'], ['users.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id')
+ )
+ _index(op.f('ix_notification_channels_id'), 'notification_channels', ['id'], unique=False)
+ _index(op.f('ix_notification_channels_user_id'), 'notification_channels', ['user_id'], unique=False)
+ _table('recovery_codes',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('user_id', sa.Integer(), nullable=False),
+ sa.Column('code_hash', sa.String(length=64), nullable=False),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('used_at', sa.DateTime(timezone=True), nullable=True),
+ sa.ForeignKeyConstraint(['user_id'], ['users.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('code_hash')
+ )
+ _index(op.f('ix_recovery_codes_id'), 'recovery_codes', ['id'], unique=False)
+ _index(op.f('ix_recovery_codes_user_id'), 'recovery_codes', ['user_id'], unique=False)
+ _table('servers',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('owner_id', sa.Integer(), nullable=False),
+ sa.Column('hostname', sa.String(length=255), nullable=False),
+ sa.Column('ip_address', sa.String(length=64), nullable=True),
+ sa.Column('os_info', sa.String(length=255), nullable=True),
+ sa.Column('cpu_cores', sa.Integer(), nullable=True),
+ sa.Column('total_ram_gb', sa.Float(), nullable=True),
+ sa.Column('check_interval_seconds', sa.Integer(), nullable=False),
+ sa.Column('collector_token_hash', sa.String(length=64), nullable=True),
+ sa.Column('machine_id', sa.String(length=64), nullable=True),
+ sa.Column('is_dashboard_host', sa.Boolean(), nullable=False),
+ sa.Column('system_info', sa.Text(), nullable=True),
+ sa.Column('name_locked', sa.Boolean(), nullable=False),
+ sa.Column('last_seen_at', sa.DateTime(timezone=True), nullable=True),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.ForeignKeyConstraint(['owner_id'], ['users.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('collector_token_hash')
+ )
+ _index(op.f('ix_servers_id'), 'servers', ['id'], unique=False)
+ _index(op.f('ix_servers_machine_id'), 'servers', ['machine_id'], unique=False)
+ _index(op.f('ix_servers_owner_id'), 'servers', ['owner_id'], unique=False)
+ _table('container_metrics',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('server_id', sa.Integer(), nullable=False),
+ sa.Column('container_id', sa.String(length=128), nullable=False),
+ sa.Column('name', sa.String(length=255), nullable=False),
+ sa.Column('image', sa.String(length=255), nullable=True),
+ sa.Column('status', sa.String(length=64), nullable=False),
+ sa.Column('restart_count', sa.Integer(), nullable=False),
+ sa.Column('cpu_pct', sa.Float(), nullable=True),
+ sa.Column('memory_used_mb', sa.Float(), nullable=True),
+ sa.Column('memory_pct', sa.Float(), nullable=True),
+ sa.Column('memory_limit_mb', sa.Float(), nullable=True),
+ sa.Column('net_rx_mb', sa.Float(), nullable=True),
+ sa.Column('net_tx_mb', sa.Float(), nullable=True),
+ sa.Column('block_read_mb', sa.Float(), nullable=True),
+ sa.Column('block_write_mb', sa.Float(), nullable=True),
+ sa.Column('pids', sa.Integer(), nullable=True),
+ sa.Column('started_at', sa.String(length=64), nullable=True),
+ sa.Column('ports', sa.String(length=512), nullable=True),
+ sa.Column('project', sa.String(length=128), nullable=True),
+ sa.Column('swap_mb', sa.Float(), nullable=True),
+ sa.Column('volumes', sa.String(length=512), nullable=True),
+ sa.Column('healthcheck_status', sa.String(length=64), nullable=True),
+ sa.Column('exit_code', sa.Integer(), nullable=True),
+ sa.Column('oom_killed', sa.Boolean(), nullable=True),
+ sa.Column('finished_at', sa.DateTime(timezone=True), nullable=True),
+ sa.Column('image_id', sa.String(length=80), nullable=True),
+ sa.Column('time', sa.DateTime(timezone=True), nullable=False),
+ sa.ForeignKeyConstraint(['server_id'], ['servers.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id')
+ )
+ _index('ix_container_metrics_server_container_time', 'container_metrics', ['server_id', 'container_id', 'time'], unique=False)
+ _index(op.f('ix_container_metrics_time'), 'container_metrics', ['time'], unique=False)
+ _table('container_rollups',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('server_id', sa.Integer(), nullable=False),
+ sa.Column('container_id', sa.String(length=128), nullable=False),
+ sa.Column('name', sa.String(length=255), nullable=False),
+ sa.Column('tier', sa.String(length=8), nullable=False),
+ sa.Column('bucket_start', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('samples', sa.Integer(), nullable=False),
+ sa.Column('down_samples', sa.Integer(), nullable=False),
+ sa.Column('avg_cpu', sa.Float(), nullable=True),
+ sa.Column('n_cpu', sa.Integer(), nullable=False),
+ sa.Column('max_cpu', sa.Float(), nullable=True),
+ sa.Column('avg_memory_mb', sa.Float(), nullable=True),
+ sa.Column('n_memory_mb', sa.Integer(), nullable=False),
+ sa.Column('max_memory_mb', sa.Float(), nullable=True),
+ sa.Column('avg_memory_pct', sa.Float(), nullable=True),
+ sa.Column('n_memory_pct', sa.Integer(), nullable=False),
+ sa.Column('max_memory_pct', sa.Float(), nullable=True),
+ sa.Column('max_memory_limit_mb', sa.Float(), nullable=True),
+ sa.Column('avg_swap_mb', sa.Float(), nullable=True),
+ sa.Column('n_swap_mb', sa.Integer(), nullable=False),
+ sa.Column('max_swap_mb', sa.Float(), nullable=True),
+ sa.Column('max_restarts', sa.Integer(), nullable=True),
+ sa.ForeignKeyConstraint(['server_id'], ['servers.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('server_id', 'container_id', 'tier', 'bucket_start')
+ )
+ _index('ix_container_rollups_tier_bucket', 'container_rollups', ['tier', 'bucket_start'], unique=False)
+ _table('firewall_states',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('server_id', sa.Integer(), nullable=False),
+ sa.Column('name', sa.String(length=16), nullable=False),
+ sa.Column('sha', sa.String(length=16), nullable=False),
+ sa.Column('rules', sa.Integer(), nullable=False),
+ sa.Column('enabled', sa.Boolean(), nullable=True),
+ sa.Column('updated_at', sa.DateTime(timezone=True), nullable=False),
+ sa.ForeignKeyConstraint(['server_id'], ['servers.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('server_id', 'name')
+ )
+ _index(op.f('ix_firewall_states_id'), 'firewall_states', ['id'], unique=False)
+ _index(op.f('ix_firewall_states_server_id'), 'firewall_states', ['server_id'], unique=False)
+ _table('host_metrics',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('server_id', sa.Integer(), nullable=False),
+ sa.Column('time', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('cpu_utilization_pct', sa.Float(), nullable=True),
+ sa.Column('load_1m', sa.Float(), nullable=True),
+ sa.Column('load_5m', sa.Float(), nullable=True),
+ sa.Column('load_15m', sa.Float(), nullable=True),
+ sa.Column('ram_used_gb', sa.Float(), nullable=True),
+ sa.Column('ram_available_gb', sa.Float(), nullable=True),
+ sa.Column('ram_pct', sa.Float(), nullable=True),
+ sa.Column('swap_pct', sa.Float(), nullable=True),
+ sa.Column('disk_usage_pct', sa.Float(), nullable=True),
+ sa.Column('disk_inodes_pct', sa.Float(), nullable=True),
+ sa.Column('net_bytes_sent_mb', sa.Float(), nullable=True),
+ sa.Column('net_bytes_recv_mb', sa.Float(), nullable=True),
+ sa.Column('system_uptime_seconds', sa.Float(), nullable=True),
+ sa.Column('boot_id', sa.String(length=64), nullable=True),
+ sa.Column('oom_kill_total', sa.Integer(), nullable=True),
+ sa.Column('kernel_tainted', sa.Integer(), nullable=True),
+ sa.Column('ntp_synced', sa.Boolean(), nullable=True),
+ sa.Column('reboot_required', sa.Boolean(), nullable=True),
+ sa.Column('last_shutdown', sa.String(length=16), nullable=True),
+ sa.Column('docker_status', sa.String(length=16), nullable=True),
+ sa.Column('docker_reason', sa.String(length=16), nullable=True),
+ sa.Column('docker_latency_ms', sa.Integer(), nullable=True),
+ sa.Column('internet_status', sa.String(length=16), nullable=True),
+ sa.Column('internet_latency_ms', sa.Integer(), nullable=True),
+ sa.Column('collect_duration_ms', sa.Integer(), nullable=True),
+ sa.ForeignKeyConstraint(['server_id'], ['servers.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id')
+ )
+ _index('ix_host_metrics_server_time', 'host_metrics', ['server_id', 'time'], unique=False)
+ _index(op.f('ix_host_metrics_time'), 'host_metrics', ['time'], unique=False)
+ _table('host_rollups',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('server_id', sa.Integer(), nullable=False),
+ sa.Column('tier', sa.String(length=8), nullable=False),
+ sa.Column('bucket_start', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('samples', sa.Integer(), nullable=False),
+ sa.Column('avg_cpu', sa.Float(), nullable=True),
+ sa.Column('n_cpu', sa.Integer(), nullable=False),
+ sa.Column('max_cpu', sa.Float(), nullable=True),
+ sa.Column('avg_load', sa.Float(), nullable=True),
+ sa.Column('n_load', sa.Integer(), nullable=False),
+ sa.Column('max_load', sa.Float(), nullable=True),
+ sa.Column('avg_ram', sa.Float(), nullable=True),
+ sa.Column('n_ram', sa.Integer(), nullable=False),
+ sa.Column('max_ram', sa.Float(), nullable=True),
+ sa.Column('avg_swap', sa.Float(), nullable=True),
+ sa.Column('n_swap', sa.Integer(), nullable=False),
+ sa.Column('max_swap', sa.Float(), nullable=True),
+ sa.Column('avg_disk', sa.Float(), nullable=True),
+ sa.Column('n_disk', sa.Integer(), nullable=False),
+ sa.Column('max_disk', sa.Float(), nullable=True),
+ sa.Column('avg_inodes', sa.Float(), nullable=True),
+ sa.Column('n_inodes', sa.Integer(), nullable=False),
+ sa.Column('max_inodes', sa.Float(), nullable=True),
+ sa.Column('avg_docker_ms', sa.Float(), nullable=True),
+ sa.Column('n_docker_ms', sa.Integer(), nullable=False),
+ sa.Column('max_docker_ms', sa.Float(), nullable=True),
+ sa.Column('avg_collect_ms', sa.Float(), nullable=True),
+ sa.Column('n_collect_ms', sa.Integer(), nullable=False),
+ sa.Column('max_collect_ms', sa.Float(), nullable=True),
+ sa.ForeignKeyConstraint(['server_id'], ['servers.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('server_id', 'tier', 'bucket_start')
+ )
+ _index('ix_host_rollups_tier_bucket', 'host_rollups', ['tier', 'bucket_start'], unique=False)
+ _table('incident_alerts',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('server_id', sa.Integer(), nullable=False),
+ sa.Column('severity', sa.String(length=16), nullable=False),
+ sa.Column('title', sa.String(length=255), nullable=False),
+ sa.Column('description', sa.Text(), nullable=False),
+ sa.Column('status', sa.String(length=32), nullable=False),
+ sa.Column('rule_key', sa.String(length=255), nullable=True),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('resolved_at', sa.DateTime(timezone=True), nullable=True),
+ sa.Column('resolution', sa.String(length=255), nullable=True),
+ sa.Column('acknowledged_at', sa.DateTime(timezone=True), nullable=True),
+ sa.Column('acknowledged_by', sa.Integer(), nullable=True),
+ sa.Column('muted_until', sa.DateTime(timezone=True), nullable=True),
+ sa.Column('notified_at', sa.DateTime(timezone=True), nullable=True),
+ sa.Column('resolve_notified_at', sa.DateTime(timezone=True), nullable=True),
+ sa.Column('reopen_after', sa.DateTime(timezone=True), nullable=True),
+ sa.ForeignKeyConstraint(['server_id'], ['servers.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id')
+ )
+ _index(op.f('ix_incident_alerts_created_at'), 'incident_alerts', ['created_at'], unique=False)
+ _index(op.f('ix_incident_alerts_id'), 'incident_alerts', ['id'], unique=False)
+ _index(op.f('ix_incident_alerts_rule_key'), 'incident_alerts', ['rule_key'], unique=False)
+ _index(op.f('ix_incident_alerts_server_id'), 'incident_alerts', ['server_id'], unique=False)
+ _table('listener_states',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('server_id', sa.Integer(), nullable=False),
+ sa.Column('proto', sa.String(length=8), nullable=False),
+ sa.Column('port', sa.Integer(), nullable=False),
+ sa.Column('scope', sa.String(length=16), nullable=False),
+ sa.Column('first_seen', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('last_seen', sa.DateTime(timezone=True), nullable=False),
+ sa.ForeignKeyConstraint(['server_id'], ['servers.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('server_id', 'proto', 'port', 'scope')
+ )
+ _index(op.f('ix_listener_states_id'), 'listener_states', ['id'], unique=False)
+ _index(op.f('ix_listener_states_server_id'), 'listener_states', ['server_id'], unique=False)
+ _table('mount_metrics',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('server_id', sa.Integer(), nullable=False),
+ sa.Column('time', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('mount', sa.String(length=255), nullable=False),
+ sa.Column('fs_type', sa.String(length=64), nullable=True),
+ sa.Column('total_gb', sa.Float(), nullable=True),
+ sa.Column('used_gb', sa.Float(), nullable=True),
+ sa.Column('used_pct', sa.Float(), nullable=True),
+ sa.Column('inodes_pct', sa.Float(), nullable=True),
+ sa.ForeignKeyConstraint(['server_id'], ['servers.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id')
+ )
+ _index('ix_mount_metrics_server_time', 'mount_metrics', ['server_id', 'time'], unique=False)
+ _index(op.f('ix_mount_metrics_time'), 'mount_metrics', ['time'], unique=False)
+ _table('mount_rollups',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('server_id', sa.Integer(), nullable=False),
+ sa.Column('mount', sa.String(length=255), nullable=False),
+ sa.Column('tier', sa.String(length=8), nullable=False),
+ sa.Column('bucket_start', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('samples', sa.Integer(), nullable=False),
+ sa.Column('avg_used_pct', sa.Float(), nullable=True),
+ sa.Column('n_used_pct', sa.Integer(), nullable=False),
+ sa.Column('max_used_pct', sa.Float(), nullable=True),
+ sa.Column('avg_used_gb', sa.Float(), nullable=True),
+ sa.Column('n_used_gb', sa.Integer(), nullable=False),
+ sa.Column('max_total_gb', sa.Float(), nullable=True),
+ sa.Column('avg_inodes_pct', sa.Float(), nullable=True),
+ sa.Column('n_inodes_pct', sa.Integer(), nullable=False),
+ sa.Column('max_inodes_pct', sa.Float(), nullable=True),
+ sa.ForeignKeyConstraint(['server_id'], ['servers.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('server_id', 'mount', 'tier', 'bucket_start')
+ )
+ _index('ix_mount_rollups_tier_bucket', 'mount_rollups', ['tier', 'bucket_start'], unique=False)
+ _table('pending_collectors',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('poll_secret_hash', sa.String(length=64), nullable=False),
+ sa.Column('pairing_code', sa.String(length=16), nullable=False),
+ sa.Column('hostname', sa.String(length=255), nullable=False),
+ sa.Column('machine_id', sa.String(length=64), nullable=True),
+ sa.Column('ip_address', sa.String(length=64), nullable=True),
+ sa.Column('os_info', sa.String(length=255), nullable=True),
+ sa.Column('status', sa.String(length=16), nullable=False),
+ sa.Column('details', sa.Text(), nullable=True),
+ sa.Column('approved_server_id', sa.Integer(), nullable=True),
+ sa.Column('join_token_id', sa.Integer(), nullable=True),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('last_poll_at', sa.DateTime(timezone=True), nullable=False),
+ sa.ForeignKeyConstraint(['approved_server_id'], ['servers.id'], ondelete='CASCADE'),
+ sa.ForeignKeyConstraint(['join_token_id'], ['join_tokens.id'], ondelete='SET NULL'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('poll_secret_hash')
+ )
+ _index(op.f('ix_pending_collectors_id'), 'pending_collectors', ['id'], unique=False)
+ _index(op.f('ix_pending_collectors_machine_id'), 'pending_collectors', ['machine_id'], unique=False)
+ _table('system_events',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('server_id', sa.Integer(), nullable=False),
+ sa.Column('time', sa.DateTime(timezone=True), nullable=False),
+ sa.Column('kind', sa.String(length=32), nullable=False),
+ sa.Column('event_key', sa.String(length=128), nullable=True),
+ sa.Column('detail', sa.String(length=512), nullable=False),
+ sa.ForeignKeyConstraint(['server_id'], ['servers.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id')
+ )
+ _index(op.f('ix_system_events_event_key'), 'system_events', ['event_key'], unique=False)
+ _index('ix_system_events_server_time', 'system_events', ['server_id', 'time'], unique=False)
+ _index(op.f('ix_system_events_time'), 'system_events', ['time'], unique=False)
+ _table('agent_reports',
+ sa.Column('id', sa.Integer(), nullable=False),
+ sa.Column('incident_id', sa.Integer(), nullable=False),
+ sa.Column('version', sa.Integer(), nullable=False),
+ sa.Column('trigger', sa.String(length=32), nullable=False),
+ sa.Column('status', sa.String(length=16), nullable=False),
+ sa.Column('model', sa.String(length=100), nullable=False),
+ sa.Column('confidence', sa.String(length=16), nullable=True),
+ sa.Column('summary', sa.Text(), nullable=True),
+ sa.Column('body', sa.Text(), nullable=True),
+ sa.Column('evidence', sa.Text(), nullable=True),
+ sa.Column('error', sa.String(length=300), nullable=True),
+ sa.Column('created_at', sa.DateTime(timezone=True), nullable=False),
+ sa.ForeignKeyConstraint(['incident_id'], ['incident_alerts.id'], ondelete='CASCADE'),
+ sa.PrimaryKeyConstraint('id'),
+ sa.UniqueConstraint('incident_id', 'version')
+ )
+ _index(op.f('ix_agent_reports_incident_id'), 'agent_reports', ['incident_id'], unique=False)
+
+
+def _bring_up_to_date(bind) -> None:
+ """The statements that ran at every start before migrations, for databases created back then. Each is safe to
+ repeat and does nothing on a database that already has the change."""
+ if bind.dialect.name == "postgresql":
+ execute = lambda statement, **values: bind.execute(sa.text(statement), values) # noqa: E731
+ for column, column_type in (("task_title", "VARCHAR(255)"), ("steps", "TEXT"), ("conversation_id", "VARCHAR(255)"), ("event_type", "VARCHAR(50)")):
+ execute(f"ALTER TABLE audit_logs ADD COLUMN IF NOT EXISTS {column} {column_type}")
+ for column, column_type in (
+ ("memory_limit_mb", "DOUBLE PRECISION"), ("net_rx_mb", "DOUBLE PRECISION"), ("net_tx_mb", "DOUBLE PRECISION"),
+ ("block_read_mb", "DOUBLE PRECISION"), ("block_write_mb", "DOUBLE PRECISION"), ("pids", "INTEGER"),
+ ("started_at", "VARCHAR(64)"), ("ports", "VARCHAR(512)"),
+ ("exit_code", "INTEGER"), ("oom_killed", "BOOLEAN"), ("finished_at", "TIMESTAMP WITH TIME ZONE"), ("image_id", "VARCHAR(80)"),
+ ("project", "VARCHAR(128)"), ("volumes", "VARCHAR(512)"), ("swap_mb", "DOUBLE PRECISION"),
+ ):
+ execute(f"ALTER TABLE container_metrics ADD COLUMN IF NOT EXISTS {column} {column_type}")
+ for column, column_type in (
+ ("boot_id", "VARCHAR(64)"), ("oom_kill_total", "INTEGER"), ("kernel_tainted", "INTEGER"),
+ ("ntp_synced", "BOOLEAN"), ("reboot_required", "BOOLEAN"), ("last_shutdown", "VARCHAR(16)"),
+ ("docker_status", "VARCHAR(16)"), ("docker_reason", "VARCHAR(16)"), ("docker_latency_ms", "INTEGER"),
+ ("internet_status", "VARCHAR(16)"), ("internet_latency_ms", "INTEGER"), ("collect_duration_ms", "INTEGER"),
+ ):
+ execute(f"ALTER TABLE host_metrics ADD COLUMN IF NOT EXISTS {column} {column_type}")
+ execute("ALTER TABLE system_events ADD COLUMN IF NOT EXISTS event_key VARCHAR(128)")
+ execute("ALTER TABLE pending_collectors ADD COLUMN IF NOT EXISTS join_token_id INTEGER REFERENCES join_tokens(id) ON DELETE SET NULL")
+ execute("ALTER TABLE pending_collectors ADD COLUMN IF NOT EXISTS details TEXT")
+ for column, column_type in (("avg_swap_mb", "DOUBLE PRECISION"), ("n_swap_mb", "INTEGER NOT NULL DEFAULT 0"), ("max_swap_mb", "DOUBLE PRECISION")):
+ execute(f"ALTER TABLE container_rollups ADD COLUMN IF NOT EXISTS {column} {column_type}")
+ execute("CREATE INDEX IF NOT EXISTS ix_system_events_event_key ON system_events (event_key)")
+ # Composite indexes that match how the metrics tables are read, then the single-column ones they made
+ # redundant. Creating an index briefly blocks writes to that table.
+ for table, name, columns in (
+ ("host_metrics", "ix_host_metrics_server_time", "server_id, time"),
+ ("mount_metrics", "ix_mount_metrics_server_time", "server_id, time"),
+ ("system_events", "ix_system_events_server_time", "server_id, time"),
+ ("container_metrics", "ix_container_metrics_server_container_time", "server_id, container_id, time"),
+ ):
+ execute(f"CREATE INDEX IF NOT EXISTS {name} ON {table} ({columns})")
+ for redundant in (
+ "ix_host_metrics_id", "ix_host_metrics_server_id", "ix_mount_metrics_id", "ix_mount_metrics_server_id",
+ "ix_system_events_id", "ix_system_events_server_id",
+ "ix_container_metrics_id", "ix_container_metrics_server_id", "ix_container_metrics_container_id",
+ ):
+ execute(f"DROP INDEX IF EXISTS {redundant}")
+ for column, column_type in (
+ ("totp_secret_enc", "TEXT"), ("totp_enabled", "BOOLEAN NOT NULL DEFAULT FALSE"), ("totp_last_step", "INTEGER"), ("password_changed_at", "TIMESTAMP WITH TIME ZONE"),
+ ("username", "VARCHAR(64)"), ("avatar_color", "VARCHAR(16)"), ("preferences", "TEXT"), ("last_seen_at", "TIMESTAMP WITH TIME ZONE"), ("agent_chat_enabled", "BOOLEAN NOT NULL DEFAULT FALSE"),
+ ):
+ execute(f"ALTER TABLE users ADD COLUMN IF NOT EXISTS {column} {column_type}")
+ for column, column_type in (
+ ("rule_key", "VARCHAR(255)"), ("resolved_at", "TIMESTAMP WITH TIME ZONE"), ("resolution", "VARCHAR(255)"), ("acknowledged_at", "TIMESTAMP WITH TIME ZONE"),
+ ("acknowledged_by", "INTEGER"), ("muted_until", "TIMESTAMP WITH TIME ZONE"), ("notified_at", "TIMESTAMP WITH TIME ZONE"),
+ ("resolve_notified_at", "TIMESTAMP WITH TIME ZONE"), ("reopen_after", "TIMESTAMP WITH TIME ZONE"),
+ ):
+ execute(f"ALTER TABLE incident_alerts ADD COLUMN IF NOT EXISTS {column} {column_type}")
+ execute("CREATE INDEX IF NOT EXISTS ix_incident_alerts_rule_key ON incident_alerts (rule_key)")
+ # The users table used to come from docker/postgres/init with its own index names; give it the models'
+ # ones. The unique index exists before the old unique constraint goes, so emails stay unique throughout.
+ execute("CREATE UNIQUE INDEX IF NOT EXISTS ix_users_email ON users (email)")
+ execute("CREATE INDEX IF NOT EXISTS ix_users_id ON users (id)")
+ execute("ALTER TABLE users DROP CONSTRAINT IF EXISTS users_email_key")
+ execute("DROP INDEX IF EXISTS idx_users_email")
+ execute("DROP INDEX IF EXISTS idx_users_role")
+ # Alerts from before the alert engine: offline alerts keep going under their rule, the rest were
+ # duplicated on every check-in with no way to resolve, so close them; the engine re-raises live ones.
+ execute("UPDATE incident_alerts SET rule_key = 'server_offline' WHERE rule_key IS NULL AND description = :description",
+ description="The collector has stopped reporting. Check that the server is up and can reach the backend.")
+ execute("UPDATE incident_alerts SET status = 'resolved', resolved_at = NOW() WHERE rule_key IS NULL AND status = 'active'")
+ # Disk and inode alerts became per mount; the old keys were for the root filesystem.
+ execute("UPDATE incident_alerts SET rule_key = rule_key || ':/' WHERE rule_key IN ('host:disk', 'host:inodes')")
+ columns = {column["name"] for column in sa.inspect(bind).get_columns("servers")}
+ for column, column_type in (("collector_token_hash", "VARCHAR(64)"), ("machine_id", "VARCHAR(64)"), ("is_dashboard_host", "BOOLEAN NOT NULL DEFAULT FALSE"),
+ ("system_info", "TEXT"), ("name_locked", "BOOLEAN NOT NULL DEFAULT FALSE")):
+ if column not in columns:
+ bind.execute(sa.text(f"ALTER TABLE servers ADD COLUMN {column} {column_type}"))
+
+
+def downgrade() -> None:
+ op.drop_index(op.f('ix_agent_reports_incident_id'), table_name='agent_reports')
+ op.drop_table('agent_reports')
+ op.drop_index(op.f('ix_system_events_time'), table_name='system_events')
+ op.drop_index('ix_system_events_server_time', table_name='system_events')
+ op.drop_index(op.f('ix_system_events_event_key'), table_name='system_events')
+ op.drop_table('system_events')
+ op.drop_index(op.f('ix_pending_collectors_machine_id'), table_name='pending_collectors')
+ op.drop_index(op.f('ix_pending_collectors_id'), table_name='pending_collectors')
+ op.drop_table('pending_collectors')
+ op.drop_index('ix_mount_rollups_tier_bucket', table_name='mount_rollups')
+ op.drop_table('mount_rollups')
+ op.drop_index(op.f('ix_mount_metrics_time'), table_name='mount_metrics')
+ op.drop_index('ix_mount_metrics_server_time', table_name='mount_metrics')
+ op.drop_table('mount_metrics')
+ op.drop_index(op.f('ix_listener_states_server_id'), table_name='listener_states')
+ op.drop_index(op.f('ix_listener_states_id'), table_name='listener_states')
+ op.drop_table('listener_states')
+ op.drop_index(op.f('ix_incident_alerts_server_id'), table_name='incident_alerts')
+ op.drop_index(op.f('ix_incident_alerts_rule_key'), table_name='incident_alerts')
+ op.drop_index(op.f('ix_incident_alerts_id'), table_name='incident_alerts')
+ op.drop_index(op.f('ix_incident_alerts_created_at'), table_name='incident_alerts')
+ op.drop_table('incident_alerts')
+ op.drop_index('ix_host_rollups_tier_bucket', table_name='host_rollups')
+ op.drop_table('host_rollups')
+ op.drop_index(op.f('ix_host_metrics_time'), table_name='host_metrics')
+ op.drop_index('ix_host_metrics_server_time', table_name='host_metrics')
+ op.drop_table('host_metrics')
+ op.drop_index(op.f('ix_firewall_states_server_id'), table_name='firewall_states')
+ op.drop_index(op.f('ix_firewall_states_id'), table_name='firewall_states')
+ op.drop_table('firewall_states')
+ op.drop_index('ix_container_rollups_tier_bucket', table_name='container_rollups')
+ op.drop_table('container_rollups')
+ op.drop_index(op.f('ix_container_metrics_time'), table_name='container_metrics')
+ op.drop_index('ix_container_metrics_server_container_time', table_name='container_metrics')
+ op.drop_table('container_metrics')
+ op.drop_index(op.f('ix_servers_owner_id'), table_name='servers')
+ op.drop_index(op.f('ix_servers_machine_id'), table_name='servers')
+ op.drop_index(op.f('ix_servers_id'), table_name='servers')
+ op.drop_table('servers')
+ op.drop_index(op.f('ix_recovery_codes_user_id'), table_name='recovery_codes')
+ op.drop_index(op.f('ix_recovery_codes_id'), table_name='recovery_codes')
+ op.drop_table('recovery_codes')
+ op.drop_index(op.f('ix_notification_channels_user_id'), table_name='notification_channels')
+ op.drop_index(op.f('ix_notification_channels_id'), table_name='notification_channels')
+ op.drop_table('notification_channels')
+ op.drop_index(op.f('ix_join_tokens_id'), table_name='join_tokens')
+ op.drop_table('join_tokens')
+ op.drop_index(op.f('ix_invites_id'), table_name='invites')
+ op.drop_table('invites')
+ op.drop_index(op.f('ix_enrollment_keys_id'), table_name='enrollment_keys')
+ op.drop_table('enrollment_keys')
+ op.drop_index(op.f('ix_check_rollups_check_id'), table_name='check_rollups')
+ op.drop_table('check_rollups')
+ op.drop_index('ix_check_results_check_time', table_name='check_results')
+ op.drop_table('check_results')
+ op.drop_index(op.f('ix_api_tokens_user_id'), table_name='api_tokens')
+ op.drop_index(op.f('ix_api_tokens_id'), table_name='api_tokens')
+ op.drop_table('api_tokens')
+ op.drop_index(op.f('ix_agent_runs_user_id'), table_name='agent_runs')
+ op.drop_index('ix_agent_runs_created', table_name='agent_runs')
+ op.drop_table('agent_runs')
+ op.drop_index(op.f('ix_users_id'), table_name='users')
+ op.drop_index(op.f('ix_users_email'), table_name='users')
+ op.drop_table('users')
+ op.drop_index(op.f('ix_conversations_user_id'), table_name='conversations')
+ op.drop_table('conversations')
+ op.drop_table('checks')
+ op.drop_index(op.f('ix_chat_messages_id'), table_name='chat_messages')
+ op.drop_index(op.f('ix_chat_messages_conversation_id'), table_name='chat_messages')
+ op.drop_table('chat_messages')
+ op.drop_index(op.f('ix_audit_logs_user_id'), table_name='audit_logs')
+ op.drop_index(op.f('ix_audit_logs_id'), table_name='audit_logs')
+ op.drop_index(op.f('ix_audit_logs_conversation_id'), table_name='audit_logs')
+ op.drop_table('audit_logs')
+ op.drop_table('app_settings')
diff --git a/tests/history/test_migrations.py b/tests/history/test_migrations.py
new file mode 100644
index 0000000..4a0df08
--- /dev/null
+++ b/tests/history/test_migrations.py
@@ -0,0 +1,90 @@
+"""Migrations (src/backend/migrations): a new database and one from before Alembic both end at the latest revision
+with exactly the schema the models describe, and running the upgrade again changes nothing.
+
+Set MIGRATION_TEST_POSTGRES_URL to an empty PostgreSQL database to run the same checks there too."""
+import os
+
+os.environ.setdefault("DATABASE_URL", "sqlite://")
+os.environ.setdefault("REDIS_URL", "redis://localhost:6379/0")
+
+import pytest
+from alembic.autogenerate import compare_metadata
+from alembic.migration import MigrationContext
+from alembic.script import ScriptDirectory
+from sqlalchemy import create_engine, inspect, text
+
+from src.backend.migrate import MIGRATIONS_DIR, alembic_config, upgrade_database
+from src.backend.models import Base
+
+POSTGRES_URL = os.getenv("MIGRATION_TEST_POSTGRES_URL")
+
+
[email protected](params=["sqlite"] + (["postgresql"] if POSTGRES_URL else []))
+def engine(request, tmp_path):
+ if request.param == "sqlite":
+ engine = create_engine(f"sqlite:///{tmp_path / 'test.db'}")
+ else:
+ engine = create_engine(POSTGRES_URL)
+ with engine.begin() as connection:
+ connection.execute(text("DROP SCHEMA public CASCADE"))
+ connection.execute(text("CREATE SCHEMA public"))
+ yield engine
+ engine.dispose()
+
+
+def _head() -> str:
+ with create_engine("sqlite://").connect() as connection:
+ return ScriptDirectory.from_config(alembic_config(connection)).get_current_head()
+
+
+def _version(engine) -> str | None:
+ with engine.connect() as connection:
+ return MigrationContext.configure(connection).get_current_revision()
+
+
+def _differences(engine) -> list:
+ with engine.connect() as connection:
+ return compare_metadata(MigrationContext.configure(connection, opts={"compare_type": True}), Base.metadata)
+
+
+def test_the_migrations_live_next_to_the_backend():
+ assert (MIGRATIONS_DIR / "versions").is_dir()
+
+
+def test_a_new_database_ends_at_the_latest_revision_with_the_models_schema(engine):
+ upgrade_database(engine)
+ assert _version(engine) == _head()
+ assert _differences(engine) == []
+
+
+def test_upgrading_twice_changes_nothing(engine):
+ upgrade_database(engine)
+ before = {table: inspect(engine).get_columns(table) for table in inspect(engine).get_table_names()}
+ upgrade_database(engine)
+ after = {table: inspect(engine).get_columns(table) for table in inspect(engine).get_table_names()}
+ assert _version(engine) == _head()
+ assert [(table, [column["name"] for column in columns]) for table, columns in before.items()] == [(table, [column["name"] for column in columns]) for table, columns in after.items()]
+
+
+def test_a_database_from_before_migrations_keeps_its_data_and_ends_at_the_latest_revision(engine):
+ Base.metadata.create_all(engine) # how the schema was made before Alembic
+ with engine.begin() as connection:
+ connection.execute(text("INSERT INTO app_settings (key, value, updated_at) VALUES ('instance_name', 'lab', CURRENT_TIMESTAMP)"))
+ upgrade_database(engine)
+ assert _version(engine) == _head()
+ assert _differences(engine) == []
+ with engine.connect() as connection:
+ assert connection.execute(text("SELECT value FROM app_settings WHERE key = 'instance_name'")).scalar() == "lab"
+
+
+def test_an_old_database_missing_tables_and_columns_is_brought_up_to_date(engine):
+ """An install that last ran an old version: no availability tables yet, and servers without the newer columns."""
+ tables = [table for name, table in Base.metadata.tables.items() if not name.startswith("check")]
+ Base.metadata.create_all(engine, tables=tables)
+ with engine.begin() as connection:
+ connection.execute(text("ALTER TABLE servers DROP COLUMN name_locked"))
+ connection.execute(text("ALTER TABLE servers DROP COLUMN system_info"))
+ upgrade_database(engine)
+ assert {"checks", "check_results", "check_rollups"} <= set(inspect(engine).get_table_names())
+ assert {"name_locked", "system_info"} <= {column["name"] for column in inspect(engine).get_columns("servers")}
+ assert _differences(engine) == []