Who this is for: the engineer who maintains the network, and the administrator who manages adapters.
Before you start: you are signed in (Getting started). Everyone in the organisation can read the network page. Adopting and retiring nodes, and revoking or rotating an adapter's credential, need an Administrator; anyone else who tries is told "Only an administrator can do this." (People and access).
The counters
Six counters sit at the top: Online, Stale, Offline, On battery, On cellular and Unclaimed. Stale, offline and unclaimed are drawn in the warning colour when they are above zero.
A node is online while its readings keep arriving, stale after three reporting intervals with nothing, and offline after twelve. A smart pole reports every five minutes by default, so it goes stale after fifteen minutes and offline after an hour.
When nothing is wrong, the page says All nodes reporting.
Nodes not reporting
Nodes not reporting lists the stale and offline nodes. Each row shows:
| Column | What it shows |
|---|---|
| Node and Site | Which node, and where. |
| Last signal | When its last reading arrived. |
| Probable cause | Power, Link or Sensor, or Being determined while it is still being worked out. |
| Evidence | One line from its last readings that points to the cause, such as the battery falling before the silence. |
The causes mean:
- Power: the battery fell before the node went quiet, or the mains supply went.
- Link: the signal fell, or the node switched to cellular.
- Sensor: one channel went flat or missing while the others kept reporting.
The same cause shows on the node's alert, so you can decide what to take to the pole before a site visit. Click a row to open the node's page. The list shows up to twenty nodes, and the card gives the total. When there are more, See all in the table opens the nodes table filtered to offline nodes; every node also has its own page.
Devices with a wrong clock
If any device reports with a clock that is wrong, Device clocks wrong lists it with how far off it is, for example "Device clock 7 min ahead" or "31 days behind". Such a node still counts as online, but its readings are marked suspect until the clock is fixed.
Adapters
An adapter is how readings and nodes reach Nebli Environment, such as a TALQ gateway or sensor boxes sending over MQTT or HTTP. For each adapter the page shows:
| Column | What it shows |
|---|---|
| Last message | When it last sent anything. |
| Rate | Messages per second. |
| Refused today | Messages turned away today, such as readings for a node the adapter does not own. |
| Suspect | Readings it sent that are marked suspect. |
| Throttled today | Whether it was slowed down today for sending too much. |
An adapter that has sent nothing for more than ten minutes is highlighted, with how long it has been quiet, for example "no message for 12 min". Its nodes then go stale and offline on their own thresholds, on the map and here.
For a TALQ gateway, View exchange shows the messages between Nebli Environment and the gateway.
Revoke or rotate a credential
Each adapter sends with its own credential. An administrator can act on it from this page:
- Rotate, for an adapter sending over MQTT or HTTP, issues a new credential. The old credential keeps working for 24 hours after you rotate, so you have time to set the new one on the devices. Other adapter kinds are revoked and recreated instead. The new secret is shown once, when it is created, and cannot be shown again; save it before you choose I have saved it.
- Revoke stops the credential. Its devices can no longer send readings, from their next message.
Both ask you to confirm first. A replaced credential shows Replaced and, during its 24 hours, the time until which it still accepts messages. You can also manage adapters under Settings › Adapters.
Delivery targets
A delivery target is a system Nebli Environment sends alerts to, such as a HikCentral video management system. For each target the page shows its last delivery and how many deliveries failed in the last 24 hours. A failed delivery is highlighted, and View deliveries shows each attempt with the receiving system's answer.
A failed delivery never affects the alert: it stays open in the console and in Alerts. While you have no delivery target, alerts show only in the console.
Unclaimed nodes
A node that reports under an id nobody has registered appears under Unclaimed nodes, with its hardware id, the adapter that reported it, and when it was first seen.
To adopt it:
- Choose Adopt on its row.
- Choose its Site. Nebli Environment suggests a Name that follows the site's last number; change it if you need to.
- Choose Adopt to confirm. The node joins the site.
A node can be left with no adapter that can speak for it, because its adapter was revoked or no longer exists. It shows No adapter with the reason, and offers Retire in place of Adopt: it can never report again. Retiring it keeps its history and frees its hardware id.
When the page cannot load
If Nebli Environment cannot be reached, the page says so with Retry. If it is only slow, it says it is taking longer than usual, with the same retry.
Next
- Alerts and rules: how rules decide when an alert opens and closes.
- Map, node page and table: every view, and exporting what you see.