Distributed Observability

StarNexus

Initializing

Loading node health, link quality, and ingress patterns.

Last sync --
Fleet Health 0 nodes 0 online
Pressure 0 degraded 0 offline
Best Node -- Waiting for scores
Weakest Link -- Waiting for probes
Ingress Hotspot -- Waiting for connection samples

Global Topology

Nodes, probe links, and live ingress overlaid in one map.

Online Degraded Offline

Event Stream

Recent incidents, status changes, and anomalies.

Active Incidents

Open, acknowledged, and suppressed issues that still need lifecycle tracking.

Control Plane

Server health, build metadata, database check, and required runtime assets.

Unknown
Version -- --
Database -- --

Statistical Radar

24h robust outlier and baseline-shift ranking across the fleet.

Building the 24h fleet radar.

Experiment View

Ground-truth fault injection evaluation from labelled experiments.

Detection -- delay --
Recovery -- delay --
False Positives -- outside labelled windows

Detector Benchmark

Five detectors replayed against the same labelled experiments with 95% bootstrap CIs.

--
Detector Detect % Mean delay (s) 95% CI FP / node-hour Events

Loading benchmark…

View head-to-head figure Detector benchmark head-to-head comparison False-positive vs detection tradeoff scatter

Scalability Benchmark

Single-server capacity from scripts/loadtest-local.sh. Static snapshot.

--
Agents Req/s p50 (ms) p95 (ms) p99 (ms) Success

Loading scalability data…

Operational Reliability

24h availability proxy, telemetry quality, and incident pressure.

Fleet Score --
Coverage --
Incidents --
Signals --

Building the reliability ledger.

Node Matrix

Sortable health, resource pressure, and score summary.

Node Status CPU Memory Down Links Score

Ingress Hotspots

Top sampled remote sources across the last 24 hours.