SafeApply pattern (validate → backup → write → reload → verify → rollback):
- HAProxy: SafeApply wraps existing validate+write+reload with backup and
post-reload health check; rolls back to .bak on failure
- dnsmasq: SafeApply validates via dnsmasq --test, backs up, atomic writes,
restarts, verifies daemon is active; rolls back on failure
- nftables: SafeApply validates via nft -c, backs up, atomic writes, applies
to kernel, verifies table loaded; rolls back on failure
Health tracking:
- Add SubsystemHealth and Health structs to reconciler
- Track per-subsystem status (ok/degraded/error) after each reconcile
- Detect recovery: previous error → current ok broadcasts recovery event
- GET /api/v1/health/reconcile endpoint exposes health state
- HAProxy tracks excluded domains as "degraded" state
SSE error events:
- Broadcast haproxy:error, dnsmasq:error on SafeApply failure
- Broadcast haproxy:recovered, dnsmasq:recovered on recovery from error
Convergence loop:
- 5-minute periodic reconcile drives system toward desired state
- Catches config drift, daemon crashes, transient failures
- Serialized by reconcile mutex — no race with event-driven reconciles
Startup:
- CheckPrerequisites verifies required (dnsmasq, haproxy) and optional
(wg, authelia, cscli, cloudflared, nft) binaries before first reconcile
- Add storage.WriteFileAtomic (temp + rename) and storage.CopyFile helpers
- Convert all 8 production config writers to atomic writes: config/state.yaml,
dnsmasq, nftables, domains, wireguard (config + secrets + peers + wg0.conf),
tunnel/cloudflared
- Add sync.Mutex to Reconciler to serialize concurrent Reconcile() calls
triggered by domain registration goroutines
- Add state.yaml backup (.bak) before every write; LoadState falls back to
backup if primary is corrupted
- Reconciler refuses to use empty config on corruption (only on first run
when no state file exists yet)
- JSON tags: fix snake_case to camelCase in dnsmasq (configFile, domainsConfigured,
lastRestart), crowdsec Machine (lastPush, lastHeartbeat), network (primaryIP,
primaryInterface). cscli raw parsing structs keep snake_case to match CLI output.
- Error wrapping: fix %v to %w in enableAuthelia for proper error chain preservation
- Naming: rename dnsmasq.ConfigGenerator to dnsmasq.Manager (matches all other packages),
rename ServiceStatus to Status in dnsmasq and haproxy (matches authelia, crowdsec, etc.)
- Logging: standardize all slog calls to use "component" key instead of message prefixes.
Affects reconcile, dnsfilter, ddns — now consistent with dnsmasq, haproxy, nftables, sse.
Adds Pi-hole-style DNS filtering using dnsmasq's native address=/ directives,
which block domains and all subdomains. Users subscribe to blocklists by URL
or upload files, with suggested lists from Hagezi, Steven Black, and OISD.
Background runner refreshes lists on a configurable interval (default 24h).
SMTP: Add host, port, username, sender, and password fields to the
Authelia configuration. Uses modern address format (submission:// for
STARTTLS on 587, smtps:// for implicit TLS on 465) with explicit TLS
server name. Falls back to filesystem notifier when SMTP is not
configured.
Fixes: Config changes now persist before attempting service restart,
so a restart failure (e.g. bad SMTP credentials) no longer prevents
saving. The specific Authelia error is extracted from the journal and
shown in the UI.
Frontend: SMTP fields added to the Authentication config card. Form
no longer continuously resets from server state while user is editing.
OIDC client edit UI added (pencil icon).
Authelia runs as a managed native service on Wild Central, providing
two integration patterns for network services:
- Forward-auth via HAProxy for apps without native SSO (Lua auth-request
script intercepts requests, redirects unauthenticated users to login portal)
- OIDC provider for apps with native support (Gitea, Grafana, etc.)
Backend: new internal/authelia/ package with service manager, config
generation, file-based user management (argon2id), and OIDC client
management. API endpoints for status, config, users CRUD, and OIDC
clients CRUD.
HAProxy: config generator extended with lua-load, auth-request
directives, and Authelia backend. Auth directives scoped to session
cookie domain — only subdomains of the auth portal's parent are
eligible for forward-auth.
Frontend: Authentication page with enable/disable, user management,
OIDC client management (add/edit/delete), and protected domains
toggles. Advanced subsystem page for raw config view. Dashboard
service card and sidebar entries.
The new icon represents Wild Central's role as a network hub —
a central node radiating connections to services. Updated favicon,
sidebar logo component, theme color, and PWA manifest name.
InstanceConfig was a ~50-field struct designed for Wild Cloud k8s instances,
but only 3 fields were ever read by dnsmasq (the sole consumer). The
reconciliation bridge constructed fake InstanceConfig objects from registered
domains just to satisfy this interface.
Replace with DNSEntry{Domain, IP} — a 2-field struct purpose-built for
dnsmasq. All domains get local=/ directives to prevent AAAA queries from
leaking to upstream DNS (Happy Eyeballs / RFC 8305 latency on LAN).
Removed dead code: InstanceConfig, NodeConfig, LoadCloudConfig,
SaveCloudConfig, DeepMerge, LoadMergedInstanceConfig, EnsureInstanceConfig,
instance path helpers, instanceLoadBalancerIP, GenerateInstanceConfig,
and all modular per-instance dnsmasq methods.
Split /api/v1/state into dedicated resource endpoints:
- /api/v1/operator, /api/v1/central-domain
- /api/v1/ddns/config, /api/v1/dns/settings
- /api/v1/dhcp/config (moved from /dnsmasq/dhcp/)
- /api/v1/nftables/config, /api/v1/haproxy/routes
Each page now talks to its own resource instead of deep-merging
a blob. Removes useConfig hook, CentralState type, and legacy
config methods.
Read the configured Central domain from state.yaml and add it as an
HTTPS origin so the web UI at e.g. https://central.payne.io works
without manually setting WILD_CENTRAL_CORS_ORIGINS.
When haproxy -c fails, parse ALERT line numbers, map to service domains
via # service: comments in generated config, exclude broken services,
and regenerate. One retry — prevents a single bad registration from
blocking all config updates.
- Add Routes model to services UI (paths, headers, IP whitelisting per route)
- Show TLS cert info per service with inline provision/renew actions
- Remove TLS Certificates section from dashboard (now on services page)
- Make gateway router port list dynamic from config + VPN state
- Add TODO for header validation in HAProxy config generation
The backend field + Public toggle already tell you where traffic
routes. Removed redundant "Routes to X on LAN, Y externally" line
and unused ddns/publicIp/backendHost vars.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The toggles (Public, Subdomains, Central TLS) already communicate
everything the table was restating. Replaced with a single routing
summary line: "Routes to X on LAN, Y externally."
Cleaner, no redundancy.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Service cards now have full inline editing:
- Backend address: editable input, updates on blur
- Public/Private: toggle switch, updates immediately
- Subdomains: toggle switch, updates immediately
- TLS (Central TLS / Passthrough): toggle switch, updates
both tls mode and backend type together
- Deregister button
Add form also uses toggle for TLS instead of dropdown.
Removed unused Badge and Select imports.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Service cards now have:
- Public/Private toggle (switches reach)
- Subdomains toggle (include *.domain)
- TLS badge (passthrough vs terminate, read-only for now)
- Inline status details (DNS, Proxy, TLS status)
- Deregister button
- Add Service form with toggles instead of dropdowns
The card speaks the user's language (public/private, subdomains on/off)
not implementation details (tcp-passthrough vs http, reach: internal).
Also updated docs/registrations.md:
- Added "User-facing concepts" section mapping API fields to toggles
- Added batch deregister endpoint
- Added backend.health field
- Cleaner examples
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Changed from single expandedDomain string to a Set<string> so
multiple cards can be open at once.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- na (passthrough, Central doesn't handle TLS): grey checkmark
- error (terminate but cert missing): red alert icon
- ok (terminate and cert exists): green checkmark
Previously passthrough showed green (wrong — Central isn't doing
anything) and missing certs showed amber (should be red — broken).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Rewrote ServicesComponent:
- Each service is its own Card (not rows in a table)
- "Add Service" button with form (domain, backend, type, reach, subdomains)
- Deregister button per service
- Wildcard services show as *.domain
- Expandable detail view with DNS/Proxy/TLS status
- Removed separate "Custom TCP Routes" section — TCP routes are just
services with type tcp-passthrough
- Added register/deregister methods to servicesApi
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Dashboard:
- Cloudflare token can be changed when already configured (was only
editable when missing/invalid)
- DDNS records can be added and removed inline (was read-only)
Services:
- Wildcard services display as *.domain instead of domain with a
separate subdomains badge — clearer at a glance
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Linter-contributed improvements:
- EnsureCentralService() method: registers Central's own domain as a
service from config, with cleanup on domain change
- Tests for EnsureCentralService: registration, idempotency, domain
change cleanup, no-domain-is-noop
- Improved reconciliation tests with Central as a registered service
- Certbot handler cleanup
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Dashboard: infrastructure health (router, cloudflare, DDNS, daemons).
Services: every registered service with full status (DNS, proxy, TLS,
DDNS) in expandable rows + custom TCP routes section.
9 sidebar items → 6. Seven old pages eliminated — their content lives
in the two new pages where it belongs.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Complete rewrite of the UI refactor plan based on the insight that
services are the entrypoint for everything. DNS, certs, proxy routes
are effects of service registrations, not independent config surfaces.
Two sidebar sections: Services (outward-facing, service-driven) and
Central (inward-facing, device config). Each service page answers
"what did Central do because of registered services?"
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Audit of current Wild Central UI navigation and page structure.
Identifies 5 problems (DNS triple redundancy, scattered Cloudflare,
jargon naming, arbitrary ordering, empty overview) and proposes a
restructured 9-item navigation with better grouping.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
dnsmasq (6 new, 30.8% → 32%):
- Empty instances, multiple instances, nil config, configured IP
- No leaked empty entries (address=//, local=//)
- Empty fields don't produce broken entries
Also fixed a bug: empty Domain/InternalDomain fields in commented-out
entries (no LB IP case) now guarded against producing # local=// and
# address=// directives.
haproxy (7 new):
- No HTTPS frontend with no routes
- HTTP-only and L4-only route scenarios
- CertsDir customization and default
- Custom TCP route ports
- Health check directive in L7 backends
services (8 new, 72.7% → 75.8%):
- Not-found errors for Get, Deregister, Update
- DeregisterBySource partial match (source AND backend)
- Overwrite preserves file count
- Empty dir and non-YAML file handling
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The reconciliation test for Central domain now verifies the HAProxy
backend uses port 15055 (the configured port) and NOT port 5055
(the default). This catches the bug where Reconcile() was called
before SetPort().
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Reconcile() was called before SetPort(), so getRunningPort() returned
the default 5055 (Wild Cloud's port) instead of 15055. Moved
Reconcile() after SetPort(port) so Central's config-driven HAProxy
route points to the actual running port.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Tests added (13 new):
- HAProxy ACL ordering: L7 before L4 wildcard, L4 exact before L4 wildcard
- HAProxy subdomains: false=exact only, true=wildcard+exact
- Reconciliation integration: HAProxy config from services, DNS config
from services, Central config domain injection, correct DNS target IPs
- dnsmasq: reach:internal has local=/, reach:public does not
- Services: idempotent registration, no duplicates in list
Bug fix:
- Run initial reconciliation on startup so Central's own domain (from
config) gets HAProxy + DNS routes even with zero registered services.
Previously, Central only got routes after the first service was
registered, causing 503 on fresh restart.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
