Dashboards¶
Edge Insights has three dashboards. Open them from Dashboards in the sidebar.
- Bus Overview: how does my system work? Frame counts, rates, CAN IDs, and timing classification.
- Bus Health: what is wrong? Anomalies, errors, foreign IDs, and congestion, with drill-down.
- Signal Explorer: decoded signal values and statistics. Needs a DBC file loaded on the interface, see Settings.
Every panel has an info icon next to its title. Hover over it for a description of what the panel shows.
Time range and refresh¶
The controls at the top of each dashboard apply to all its panels:
- Time range: select the range to choose a preset, e.g. the last 15 minutes, or set explicit start and end times.
- Zoom out: the magnifier button next to the time range doubles the range.
- Drag to zoom: drag across a chart to zoom into that time range.
- Filters: e.g. CAN Interface, Anomaly Type, and CAN ID, depending on the dashboard.
- Refresh: loads new data now. The list next to it sets how often the dashboard refreshes on its own, or turns auto-refresh off.
- Export: opens Export Log with the interface and time range of the dashboard.
A panel with no data in the time range offers Zoom to data when newer or older data exists. It shows the 15 minutes up to the newest data.
Each chart has two buttons in its top-right corner. The table button shows the data of the chart as a table. Select it again to return to the chart. The expand button shows the panel at full size.
Bus Overview¶
- At a glance: frames in range, active CAN IDs in range, bus utilization, and anomalies in range.
- Bus utilization and CAN frame rate by ID: bus load over time, and the frame rate of each CAN ID.
- All CAN IDs timing summary: one row per CAN ID and interface, with its class, confidence, sample count, average interval, jitter, and anomaly count. The same CAN ID seen on two interfaces is listed and classified separately for each one.
- Classification distribution: the share of periodic, aperiodic, and burst CAN IDs.
Bus Health¶
- At a glance: anomalies in range, timing anomalies, congestion windows, bus errors, and foreign ID events. The counts follow the time range, so two readings taken at different times are comparable only if they cover the same range.
- Anomaly timeline: anomalies over time, split by type: timing, congestion, foreign ID, and bus error.
- Anomaly list: every anomaly in range, with timestamp, type, CAN ID, interface, a summary, and a count. Timing anomalies show as Late, Early, or Gap rows. Filter by CAN interface, anomaly type, or CAN ID with the controls at the top.
Anomaly drill-down¶
Select a row in the Anomaly list on Bus Health to open a deep-dive dashboard for that anomaly. Which dashboard opens depends on the type in the row.
| Type | Opens | Filtered to |
|---|---|---|
| Late, Early, Gap | Timing Anomaly Deep Dive | The CAN ID in the row |
| Congestion | Congestion Deep Dive | The interface |
| Foreign ID | Foreign ID Deep Dive | The interface |
| Bus Error | CAN Errors Deep Dive | The interface |
Only a timing anomaly belongs to a single CAN ID. The other three describe the state of the bus, so they open filtered to the interface. Every deep dive opens on a time range around the event, with about 30 seconds of context on each side. Change it with the time range control. The back arrow next to the title, or the browser's Back button, returns to Bus Health with the time range you had.
Timing¶
The Interval timeline panel plots the measured interval and one reference line for each point in time.
- Measured (window avg): the average time between frames for that CAN ID, over one aggregation window.
- Baseline (frozen): the interval stored when the learn phase ended. The analyzer scores against this value from that moment.
- Expected (live): the median of recent intervals. The analyzer scores against this value in direct mode and before a baseline is frozen.
Edge Insights learns both reference values from the bus. Neither is the cycle time declared in the DBC file. Anomalies are scored against the reference line, not against the measured average. Red points mark the individual intervals that fell outside the threshold.
The lines answer different questions. If the measured line sits on the reference line, the message runs at its normal cadence. If the measured line drifts away from Baseline (frozen) while few frames are flagged, the cadence has shifted, but each frame is still inside the allowed band. This is an early sign of a timing problem. If the measured line stays on the reference line while red points appear, the average cadence is normal and individual frames are jittering. Expected (live) follows the traffic, so a slow shift in cadence is easier to see against Baseline (frozen).
The Anomaly details table lists every flagged frame with its measured interval, the baseline it was scored against, and the deviation between them.
Congestion¶
A microburst is a short window, well under a second, where bus utilization crossed the congestion threshold. Set the threshold in Settings.
The dashboard shows the same traffic at two time scales, and the difference between them is the point.
- Sub-second (why it triggered): the Micro-burst timeline plots one point per event, at the utilization measured during that short window. These are individual events, not a continuous signal.
- Per-second (context): the Bus utilization panel plots the per-second average over the same range.
A burst that saturates the bus for 200 ms stands out in the first panel and is almost invisible in the second, because the per-second average spreads it over five times its duration. Read the two together to tell an isolated burst from sustained pressure. The Avg bus utilization and Peak utilization numbers at the top make the same comparison in a single figure each.
The Recent events table lists every microburst with its utilization, frame count, payload bytes, and the threshold that applied at the time. The threshold is recorded per event, so events from before a settings change still show the value they were judged against.
Foreign ID¶
A foreign ID is a CAN ID that is not defined in the loaded DBC file, not learned into the baseline, and not listed under Known CAN IDs. It can mean a new or misconfigured ECU, a device left connected from an earlier test, or traffic that does not belong on the bus.
The Frame count timeline plots one series per foreign ID, so a steady talker and a one-time appearance look different at a glance. The Event details table lists every event with its frame count and the timestamp the ID was first seen.
The drill-down opens with every foreign ID on the interface selected. Narrow it with the CAN ID filter at the top when one ID is of interest.
Before treating a foreign ID as a fault, check that it is not simply missing from your configuration. Adding it under Known CAN IDs stops the anomalies.
Bus error¶
These are errors reported by the CAN controller itself, not anomalies derived from traffic patterns. A Bus Error anomaly fires when the number of error frames in a window exceeds the error rate threshold, both set in Settings.
Error rate over time breaks the errors down by class: TX timeout, lost arbitration, controller, protocol, transceiver, no ACK, bus off, bus error, restarted, and error counter. The class points at the layer to investigate. No ACK usually means nobody is listening, for example a single node on the bus or a wiring fault. Lost arbitration is normal in small numbers on a busy bus. Bus off means the controller took itself off the bus after too many errors.
Two details when reading the stacked chart. A single error frame that carries several class bits is counted once in the totals but appears in each matching series, so the stacked total can be higher than the raw frame count. Bus off and restarted are state transitions rather than running counters, so they appear as isolated marks rather than a rate.
The Recent threshold events table lists the windows that crossed the threshold, each with its own class breakdown.
Signal Explorer¶
Signal Explorer plots decoded signals. Available IDs and Available signals count what was decoded in the time range.
Select Add graph to add a graph, then pick one or more signals for it. Below each graph, a table lists every signal in it with its CAN ID, sample count, minimum, maximum, mean, and last value.
Each graph has buttons in its top-right corner to show the data as a table, move the graph by dragging, edit its signals, remove it, and show it at full size.
Signal Explorer draws every decoded sample, so a long time range can take a moment to load.






