Some deployments never ran migration 0007 that created the
linking_tokens table. The camelCase rename migration (0021) now
creates the table with old column names if missing before renaming.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
SQLite was too slow for analytical aggregations on traffic_events and
waf_events (millions of rows, GROUP BY, COUNT DISTINCT). ClickHouse is
a columnar OLAP database purpose-built for this workload.
- Add ClickHouse container to Docker Compose with health check
- Create src/lib/clickhouse/client.ts with singleton client, table DDL,
insert helpers, and all analytics query functions
- Update log-parser.ts and waf-log-parser.ts to write to ClickHouse
- Remove purgeOldEntries — ClickHouse TTL handles 90-day retention
- Rewrite analytics-db.ts and waf-events.ts to query ClickHouse
- Remove trafficEvents/wafEvents from SQLite schema, add migration
- CLICKHOUSE_PASSWORD is required (no hardcoded default)
- Update .env.example, README, and test infrastructure
API response shapes are unchanged — no frontend modifications needed.
Parse state (file offsets) remains in SQLite.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace client-controlled redirectUri with server-side redirect intents.
The portal server component validates the ?rd= hostname against
isForwardAuthDomain, stores the URI in a new forward_auth_redirect_intents
table, and passes only an opaque rid (128-bit random, SHA-256 hashed) to
the client. Login endpoints consume the intent atomically (one-time use,
10-minute TTL) and retrieve the stored URI — the client never sends the
redirect URL to any API endpoint.
Security properties:
- Redirect URI is never client-controlled in API requests
- rid is 128-bit random, stored as SHA-256 hash (not reversible from DB)
- Atomic one-time consumption prevents replay
- 10-minute TTL limits attack window for OAuth round-trip
- Immediate deletion after consumption
- Expired intents cleaned up opportunistically
- Hostname validated against registered forward-auth domains before storage
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
CPM can now act as its own forward auth provider for proxied sites.
Users authenticate at a login portal (credentials or OAuth) and Caddy
gates access via a verify subrequest, eliminating the need for external
IdPs like Authentik.
Key components:
- Forward auth flow: verify endpoint, exchange code callback, login portal
- User groups with membership management
- Per-proxy-host access control (users and/or groups)
- Caddy config generation for forward_auth handler + callback route
- OAuth and credential login on the portal page
- Admin UI: groups page, inline user/group assignment in proxy host form
- REST API: /api/v1/groups, /api/v1/forward-auth-sessions, per-host access
- Integration tests for groups and forward auth schema
Also fixes mTLS E2E test selectors broken by the RBAC refactor.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- New l4_proxy_hosts table and Drizzle migration (0015)
- Full CRUD model layer with validation, audit logging, and Caddy config
generation (buildL4Servers integrating into buildCaddyDocument)
- Server actions, paginated list page, create/edit/delete dialogs
- L4 port manager sidecar (docker/l4-port-manager) that auto-recreates
the caddy container when port mappings change via a trigger file
- Auto-detects Docker Compose project name from caddy container labels
- Supports both named-volume and bind-mount (COMPOSE_HOST_DIR) deployments
- getL4PortsStatus simplified: status file is sole source of truth,
trigger files deleted after processing to prevent stuck 'Waiting' banner
- Navigation entry added (CableIcon)
- Tests: unit (entrypoint.sh invariants + validation), integration (ports
lifecycle + caddy config), E2E (CRUD + functional routing)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
_journal.json must list every migration for the Drizzle migrator to pick
it up at runtime. 0014_waf_blocked was missing, so the blocked column was
never added to waf_events.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Drizzle's better-sqlite3 migrator splits SQL files on --> statement-breakpoint
before running each statement. Without it, multi-statement files fail with
"The supplied SQL string contains more than one statement".
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- DetectionOnly mode: add SecAction to set anomaly score thresholds to
9999999 so rule 949110/980130 never fires; works around coraza-caddy
bug where is_interrupted=true still causes a 403 in detection mode
- Switch SecAuditEngine back to On (from RelevantOnly) so DetectionOnly
hits are captured, now safe because body parts are excluded
- SecAuditLogParts: ABIJDEFHZ → ABFHZ, dropping request body (I),
multipart files (J), intermediate response headers (D), and response
body (E) — prevents multi-MB payloads being written to audit log
- Parser: store both blocked and detected events; filter on rule matched
OR is_interrupted instead of is_interrupted only
- Add blocked column to waf_events (migration 0014); existing rows
default to blocked=true
- WAF Events UI: Blocked/Detected chip in table and drawer header
- Fix misleading help text that said to use Detection Only to observe
traffic before blocking
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- New `ca_certificates` table for reusable CA certs (migration 0011)
- CA cert CRUD model, server actions, and UI dialogs
- Proxy host create/edit dialogs include mTLS toggle + CA cert selection
- Caddy config generates `client_authentication` TLS policy blocks with
`require_and_verify` mode for hosts with mTLS enabled
- CA certs sync to slave instances via instance-sync payload
- Certificates page shows CA Certificates section
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Parse Caddy access logs every 30s into traffic_events SQLite table
- GeoIP country lookup via maxmind (GeoLite2-Country.mmdb)
- 90-day retention with automatic purge
- Analytics page with interval (24h/7d/30d) and per-host filtering:
- Stats cards: total requests, unique IPs, blocked count, block rate
- Requests-over-time area chart (ApexCharts)
- SVG world choropleth map (d3-geo + topojson-client, React 19 compatible)
- Top countries table with flag emojis
- HTTP protocol donut chart
- Top user agents horizontal bar chart
- Recent blocked requests table with pagination
- Traffic (24h) summary card on Overview page linking to analytics
- 7 authenticated API routes under /api/analytics/
- Share caddy-logs volume with web container (read-only)
- group_add caddy GID to web container for log file read access
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Drizzle's better-sqlite3 migrator splits on '--> statement-breakpoint'.
Without it, the entire file is passed to db.prepare() as a single
statement, which better-sqlite3 rejects with 'more than one statement'.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Change providerSubjectIdx from index to uniqueIndex in schema.ts to
prevent multiple users sharing the same (provider, subject) pair,
which caused non-deterministic sign-in resolution via findFirst.
- Add migration 0008_unique_provider_subject.sql: DROP the existing
non-unique index and CREATE UNIQUE INDEX in its place.
- Validate INSTANCE_SYNC_MAX_BYTES env var in sync route: fall back to
10 MB default when the value is non-numeric (e.g. 'off') or
non-positive, preventing NaN comparisons that silently disabled the
size limit.
- Return a generic error message to callers on applySyncPayload /
applyCaddyConfig failure instead of leaking the raw error string;
the original message is still stored internally via setSlaveLastSync.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
