* fix(ci): pin gitleaks-action to valid commit after upstream force-push
Upstream deleted dcedce43 (force-push), so every ci.yml run fails at
workflow-parse with 0 jobs. The dependabot actions-group bump (#30)
contains this fix but cannot merge whole: its docker/* actions require
node24, which GitHub runners do not support yet. Pin gitleaks alone
(ff98106e is node20).
* fix(ci): downgrade v7 actions to node20-compatible versions
actions/checkout@v7, setup-node@v7, setup-python@v7 require node24,
which GitHub-hosted runners do not support yet — every ci.yml run has
failed at workflow-parse (0 jobs) since the Aug 3 v7 bump. Pin to the
latest node20 versions (checkout v4, setup-node v4, setup-python v5).
Also pins gitleaks-action to a valid commit (upstream force-pushed away
dcedce43).
* ci: noop retrigger
* fix(ci): escape literal ${{ in bash string so GitHub parser accepts workflow
The 'Validate tracked build and Compose inventory' step matched literal
${{ inside run-block strings, which GitHub's expression parser reads as
the start of an expression (${{'* is invalid) — the workflow fails at
parse time with 0 jobs. Split the literal as '$''{{' (bash concatenates
adjacent quoted strings at runtime; the source no longer contains a
contiguous ${{).
* fix(ci): resolve all three pre-existing ukmesh CI failures
Backend (packetBatchWriteCoalesce):
- 045: restore nodes_public_visibility_generation trigger — 015 is
superseded by 044 on fresh DBs so the trigger was never created,
leaving generation/visibility_generation out of sync and public
packet reads returning empty.
- 046: restore packet privacy classification — the 042 fence design
assumes sync_private_node_prefixes rewrites packets on privacy
change, but 026's version only maintains prefixes. Node flips to
private left old packets visible; direct SQL inserts were never
classified. Restore the packet rewrite in the sync trigger and add
a BEFORE INSERT trigger mirroring the batch path's is_private /
visibility_ok computation.
- packetBatch.integration.test.ts: exclude the prefix-cache refresh
query from statementCount (matches unit-test convention).
Frontend e2e:
- public.spec.ts: assert the TopologyMap component's actual labels
('Geographic repeater topology map', '2 mapped repeaters · 2
observed relationships') instead of the pre-map SVG graph labels.
Workers (pip install):
- viewshed-worker: bump numpy 1.26.4→2.3.5, scipy 1.13.1→1.16.3,
psycopg2-binary 2.9.10→2.9.11 (cp314 wheels); the base gdal image
ships Python 3.14 so the old pins had no wheels. Inherit shapely
from the image's python3-shapely apt package (no cp314 wheel
exists; source builds are GEOS 3.14-incompatible).
* fix(ci): restore data-plane services for compose validation and smoke test
Commit 8c5e1c8 split the long-lived data-plane services (timescaledb,
mosquitto, redis, mosquitto-reloader) into the external meshcore-infra
project on the live host, but the Workers-and-Compose CI job still
asserted their presence in the app compose project (max_worker_processes
check, inventory check, smoke-test execs). CI had been broken since
Aug 3 so the mismatch was never caught.
Add docker-compose.ci.yml — a CI-only overlay restoring the four
data-plane services from their pre-split definitions — and point every
compose invocation in the Workers-and-Compose job at
'-f docker-compose.yml -f docker-compose.ci.yml'.
Validated locally: merged config parses, the jq assertions and the full
inventory (db-migrate backend app-ukmesh website-ukmesh website-dev
mesh-health-check mosquitto-reloader link-worker link-backfill-worker
hopreach timescaledb mosquitto) all pass.
* fix(ci): include dev profile in compose inventory check
website-dev is a dev-profile service; the inventory gate compared the
base-config service list, so the merged stack never matched. This gate
has been red since the workflow was first written (CI parse-broken from
Aug 3 until the gitleaks pin landed).
* fix(ci): db-migrate must wait for timescaledb health
The Aug 9 data-plane split removed db-migrate's depends_on along with
the timescaledb service definition, so the migration runner raced the
fresh database init in the empty-volume smoke test and failed on the
not-yet-created base schema. Restore the pre-split condition:
timescaledb service_healthy.
* fix(ci): restore internal TCP MQTT listener for smoke test + exporter
Commit 40e843f dropped the 1883 TCP listener, breaking the CI smoke
(mosquitto_pub targets 1883) and the Prometheus mosquitto exporter.
The listener is internal-only: compose never publishes 1883 on the
host, and the password/ACL policy still applies.
---------
Co-authored-by: gadgethd <111318106+gadgethd@users.noreply.github.com>
* Fix map node freshness consistency
* Harden output, ingest, caches, and WebSocket limits
* Enforce public visibility across derived data
* Harden proxy and operator deployment boundary
* Make owner grants authoritative and reconcile ACLs safely
* Bound path, spam, and statistics analysis
* Make link and coverage jobs crash-safe
* Implement strategic security remediation
* Fix production cutover configuration
* Fix disabled viewshed worker health signal
* Serve stale stats during background refresh
* Retain stale stats through refresh windows
* Bound analytics work to protect ingestion
* Prioritize summary warmup over chart scans
* Throttle path history rebuilds
* Bound path history result memory
* Stream path history aggregation
* Give bounded path rebuild one CPU
* Serve stale charts during bounded refresh
* Prioritize startup stats before chart scans
* Bound path history segment cardinality
* Pin path rebuild context to privacy generation
* Self-host original frontend fonts
* Allow bounded path rebuild to complete
* Improve live map UI and low-latency group feed
- Dock node details on the right with selection highlight and collapsible layers
- Add node legend, 24h activity sparkline, copy-link, and layout/overlap fixes
- Keep all repeaters visible during Live Path focus
- Send GroupText feed packets immediately over WebSocket (no batch delay)
- Cache expensive stats/observer activity more aggressively to protect ingest
- Remove stale local planning/audit markdown from the tree
* fix(ci): supply OPERATOR_SITE_TOKEN for compose validation
Workers/Compose CI failed because docker-compose requires
OPERATOR_SITE_TOKEN. Add CI placeholders for that and MQTT_PASSWORD.
Consolidates the stacked backend, privacy, network-intelligence, frontend, operations, mobile, and owner-cache changes after resolving main conflicts and passing the full CI suite.
Use ${TERRAIN_TILES_PATH:-./terrain-tiles} so the path is configurable
per-host without leaking machine-specific paths in the repo.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Multi-node LOS: multiple repeaters can have LOS active simultaneously,
each auto-expiring after 15 seconds; LOS persists after popup closes
- Custom LOS tool: desktop-only map button to sample terrain between two
points (click map or repeater); renders per-segment green/clear or
red/obstructed based on terrain height; prohibited nodes are excluded
- Terrain exaggeration: deck.gl altitudes multiplied by exaggeration factor
(2×) to match MapLibre's visual terrain mesh
- LOS line rendering: bloom+core PathLayer tube effect with ScatterplotLayer
endpoint markers; popup z-index fix so it renders above deck.gl canvas
- Terrain sampler: client-side Terrarium PNG tile decoder for elevation sampling
- DB schema: terrain_profile_json JSONB column added to node_links
- Nginx: terrain tile serving block added
- Docker: terrain tiles volume mount added
Excluded: PacketDetailPanel, UKFeedPage, styles/globals.css (unrelated local
changes), backfill_profiles.py (server-side utility, not intended for repo)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Lazy path resolver: geographic hop reconstruction from path_hashes without
pre-training data; fixes observer self-hash trimming bug
- ACL watcher worker: auto-provisions Mosquitto ACL entries when PUBLISH is
denied for a known user; eliminates the recurring gnomemoe ACL removal issue
- connectionMonitor: increase historical scan to 50 MB; fix OOM by streaming
lines instead of buffering; fixes missed backfill on restart
- Feed page: replace path-beta with lazy paths; add clickable node markers
showing name and public key; observer column nowrap; summary text unclipped
- PacketFeed: GRP-only message store protected from ADV eviction; scroll to
bottom on new message; 24-hour message backfill on initial load
- getRecentMessages: separate DB query for last 50 GRP messages last 24h
- getPacketDetail: new endpoint for full packet detail including observations
- Grafana/Loki/Promtail observability stack with Docker log collection
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- New aclManager.ts rewrites a user's ACL block (all linked nodes) and
sends SIGHUP to the Mosquitto container whenever autoLinkOwnerNodeIds
runs at owner login, eliminating the need to manually edit the ACL
file when a user connects a second node.
- requireOwnerSession now resolves node IDs from owner_account_nodes
on every request instead of using the frozen session cookie list,
so manually- or auto-linked nodes are visible without re-login.
- docker-compose: mount mosquitto config dir (rw) and Docker socket
into backend so it can update the ACL file and signal Mosquitto.
- Add dockerode dependency for container SIGHUP signalling.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- narrow owner last-hop telemetry to the selected MQTT node and cache it per node with hourly background refresh
- tighten repeater-only last-hop classification and clean up the owner chart selector/filtering
- fix owner direct-sender map basemap, static behavior, and viewport fitting/centering
- allow bare ukmesh.com and teessidemesh.com host handling in backend/frontend config
Owner portal fixes:
- Fix session node ID case mismatch: decryptOwnerSession was lowercasing
node IDs but /owner/live uppercased the ?nodeId= param, so the find()
never matched — users always got "Node is not owned by this session"
- Revert sameSite strict → lax (broke sessions after March 18 deploy)
- Remove LOWER() wrappers from all owner live SQL (node IDs are uppercase)
- Add Fireball's second node (TS17 Yagi North) to owner_account_nodes
Tile worker:
- Server-side PNG tile rendering via @napi-rs/canvas
- Tile snapshot cache, Redis-backed tile cache, tile worker service
- /api/tiles/nodes/:z/:x/:y.png endpoint with rate limiter
- UK tile TTL logic
Cross-network connectivity:
- Add 7-day hourly history series to /cross-network-connectivity response
- Cache result for 60s to avoid repeated join-heavy query
- Fix query to use literal interval instead of interpolated variable
Other:
- Remove LOWER() from packet/node SQL throughout (IDs stored uppercase)
- Various performance and cleanup improvements
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
RADIO_BOT_URL is a private integration — strip the default service name
from docker-compose and remove it from the README env table.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- New www.teessidemesh.com single-page dashboard (TeessieDashboard.tsx)
with amber/charcoal theme, connectivity status, repeater cards with
7-day battery sparklines, MQTT observer grid, and decoded packet feed
- Backend: cross-network connectivity endpoint (hop-count packet matching),
radio-stats and radio-history proxy endpoints, observer-activity endpoint,
nodesDay stat field
- docker-compose: website-teesside service on port 3002
- Path-beta: worker pool, resolve cache, and pool refactor for performance
- Connection monitor, WS server and MQTT client improvements
- Mosquitto config, viewshed worker, and frontend UI polish
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Drop duplicate MQTT packets when the same packet hash arrives at the
same observer with the same hop count within 120s. Increase TimescaleDB
max_connections to 300 to prevent connection exhaustion on startup.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- move beta path resolution onto the backend with packet-level diagnostics and live API access
- add path simulation worker plumbing, health/schema updates, and compose wiring
- update app overlays to fetch server-computed beta paths with animated rendering and request caching
- refine mobile controls so map layers collapse behind a toggle and compact the links legend
www.ukmesh.com (website-ukmesh container, port 3004):
- UKLayout, UKHomePage, UKInstallPage, UKMqttPage components
- VITE_SITE=ukmesh build arg routes to UK pages in main.tsx
- Stats fetched via relative /api path (proxied by Nginx)
app.ukmesh.com (app-ukmesh container, port 3003):
- Same dashboard build with VITE_APP_HOSTNAME=app.ukmesh.com
- VITE_NETWORK=ukmesh for future API filtering
MQTT:
- node-ukmesh Mosquitto user for ukmesh observer nodes
- ACL updated: backend reads meshcore/# and ukmesh/#
- ukmesh nodes publish to ukmesh/{IATA}/{PUBKEY}/packets
Cloudflare routes to add in Zero Trust:
www.ukmesh.com -> http://website-ukmesh:80
app.ukmesh.com -> http://app-ukmesh:80
mqtt.ukmesh.com -> http://mosquitto:9001
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Splits the monolithic app container into:
- backend: Node.js handling MQTT, WebSocket, API, DB (no static files)
- app: Nginx serving the React SPA, proxying /api and /ws to backend
This means future dashboards (app.ukmesh.com etc.) can be added as
additional Nginx containers pointing at the same backend with no
duplicate calculations or DB connections.
Also updates nginx.website.conf to proxy to backend instead of app,
and bumps website host port from 3001 to 3002 to avoid conflict.
NOTE: Cloudflare Zero Trust tunnel route for app.teessidemesh.com
must be updated from http://app:3000 to http://app:80.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
gadgethd