Explorer and exporter were scanning the full Parquet lake every 1–5s
(~2000 files, 200MB+), driving ~2.2 CPU cores. Limit metrics to the
last N flights, cache DuckDB views and the partition tree, slow live
polls to 2s, and keep drones alive after seal to stop restart churn.
Add node-exporter, scan-duration metrics, and a Swarm Platform Grafana
dashboard for node CPU/memory and scan health.
Use the real hive column name (`drone`) in the prototype's live query
and pin the Grafana datasource uid to `swarm-prom` so provisioned
dashboards resolve their Prometheus source in k3d.
Non-root drone pods need write access on the hostPath lake. Ansible now
creates var/t1 and var/t3 owned by uid 10001 instead of relying on 0777
alone after legacy root-owned parquet trees.
Add swarm uid 10001 in both build stages so Trivy config scan passes
and containers do not run as root. Mounted /data volumes should be
world-writable or owned by uid 10001 (var/t1 from Ansible is 0777).
Document why WireGuard and SSH are complementary layers, add Flux
boundaries for the ground segment, shared var/ data paths in docs/06,
and a contributor guide with the required test-first workflow.
Ansible provisions the k3d cluster and Flux controllers; Terraform
modules deploy the simulated fleet and ground warehouse. Compose and
k3d share var/t1 (live lake) and var/t3 (warehouse). The prototype
gains a live mode fed by the explorer read-only SQL API.
Unit tests cover the SQL gate, broadcast codec, Hive writer layout,
and flight kinematics. CI runs pytest first; the simulator image build
runs tests in a dedicated stage before the final layer.
Grafana was checking grafana.com for updates and sending usage
reporting by default - the only outbound traffic in the whole stack.
Verified empty latestVersion and no established external connections
from any container after the change.
The field was a fixed square, leaving wide monitors with dead side
margins. Field width now follows the window aspect (capped at 2.2:1,
quantized to avoid rebuilds on minor resizes); routes, obstacles and
transit spawns stretch with it, and the grid keeps 100 m cells.
Mobile obstacles used to bounce until they wedged into far corners, and
a single perimeter route let the mesh stretch into segments that stopped
exchanging data. Even drones now fly the perimeter while odd drones fly
an X through the center, keeping bridge links alive, and transit objects
enter at one edge, cross the area, and respawn at a fresh edge.
Gitea Actions: compile check, virtual-drone smoke flight validated with
DuckDB, read-only SQL gate tests, Trivy vuln and misconfig scans. Pushes
to main cut a semantic version from the conventional-commit type and
publish a release with a fleet release manifest and source bundle.
Single-page view over the Parquet lake: live partition tree with
rolled-up sizes and a SQL console gated by the same read-only rules
as the peer query channel. Ships in the monitoring profile.
Bench and flight share one query channel, so debugging exercises the
production path daily. Grafana/Prometheus/exporter host ports become
env-overridable for shared hosts.
Record the full reasoning path (GraphQL service, Arrow Flight, raw SSH,
forced-command synthesis) and specify read-only enforcement as five
fail-safe layers: key-per-operation, SELECT-only statement gate,
read-only OS user, engine hardening, resource caps. Promote 'SQL is
the contract' to a top-level design principle.
Monitoring compose profile: dependency-free Prometheus exporter that
scans the Parquet output with DuckDB (rows, detections, pose frames,
battery, RSSI, link distance, bytes on disk), Prometheus scrape config
and a provisioned Grafana dashboard. Demonstrates the observability
doctrine: fleet statistics derived from the data platform itself.
Replace object-store replication with rsync over persistent multiplexed
SSH (ed25519, pinned known_hosts, forced commands as the only remote
API); demote GraphQL to the ground integration standard. Add DuckDB
spatial extension for separation audits, perimeter predicates and
trajectory analysis in the local planar frame.
Top-down SVG view: patrol flight with separation and obstacle avoidance,
mesh links appearing within radio range, pose-broadcast flashes, bulk
sync transfers with live rate and cumulative up/down counters, channel
hopping and battery per drone. Pure client-side simulation (Vite/React/TS).
One process per drone: seeded patrol kinematics, sensor streams at
realistic rates, detection events, hot current/ blocks sealed into
Hive-partitioned Parquet, and the compact UDP state broadcast with
peer RSSI derivation. Scales via docker compose --scale.
eSlider