Connection Management
Table of Contents
- Overview
- Screen Layout — Three Submenus
- Protocol Screen (
/connect/index) - Status Screen (
/connect/status) - Data Source Detail Screen (
/connect/view) - Edge Screen (
/edge/index) - Edge Detail Screen (
/edge/view)- Page Header and Refresh
- Left — Four Panels: Basic Info, Host, Network, Security
- Right ① System Resource Gauges + App Restart
- Right ② PLC / Collection Statistics (8-Cell stat-box)
- Right ③ OPC Status + 5 Summary Chips + Start/Stop Collection
- Right ④ Docker Containers (Remote Deploy, Start, Stop, Logs, Delete)
- Safe Procedures for System Reboot and App Restart
- Usage Scenarios
- Frequently Asked Questions
- Related Screens
Overview
The Connectivity group is where you register and inspect connections used to collect data from external systems (field PLCs, OPC servers, files, external DBs, MQTT/Kafka topics, and so on). The menu consists of the following three sub-screens.
| Submenu | Internal URL | Purpose |
|---|---|---|
| Protocol | /connect/index | Register a new connection — entry point for adding each data source type |
| Connection Status | /connect/status | Monitor communication status, latency, and validation failures of live connections |
| Edge | /edge/index | System status and PLC collection statistics of field edge devices |
Screen Layout — Three Submenus
| Screen | Primary user | One-line summary |
|---|---|---|
| Protocol | System administrator | Add a new data source (Edge/OPC/PLC/Modbus/external DB/file·API) |
| Connection Status | Operator | Trends for normal/delayed/abnormal connections, plus 24-hour disconnection, latency, and validation failure history |
| Edge | Operator | Edge OS status (CPU/memory/disk/network) + PLC collection statistics |
The connection state names used in this document and on screen are Connected · Connection Delayed · Connection Abnormal · Disconnected · Not Observed
(five in total). Their code values, in order, are OK · WARN · ERROR · DISCONNECTED · UNKNOWN, and the
definitions are in Status Code Definitions.
Some older screens, such as the asset tree legend, still label OK as "Normal Connection" — it is the same state.
Protocol Screen (/connect/index)
An information box appears at the top of the screen.
🌐 Connections to industrial protocols and other data sources [Recommended] Integration with the PlantPulse Edge gateway is seamless and requires no configuration. You can collect data by connecting to widely used industrial OPC, PLC, and Modbus protocols as well as databases. In addition, the supplementary client API lets you collect from MQTT, STOMP, KAFKA, and HTTP. (OPC connects automatically; everything else uses the Edge gateway or collector matching the protocol.)
Data Source Categories
The page is divided into large sections by category, and each section shows a grid of icon cards for the data source types (each with a Connect link).
| Section | Cards shown | Additional link |
|---|---|---|
Edge Gateway / File / API [N] | Edge gateway, MQTT, REST API, Excel, CSV (5 types) | /connect/file/add |
OPC [N] | OPC standard data sources (automatic connection) | /connect/opc/index list |
PLC [N] | PLC communication (protocol by manufacturer) | /connect/plc/index list |
Modbus [N] | Modbus TCP/RTU | /connect/modbus/index list |
Database [N] | External DB polling | /connect/database/index list |
The [N] next to each section label is the number of connections currently registered. The "View list" link on the right takes you to the list screen for that category.
OPC List Screen (/connect/opc/index)
Reached from the OPC row of the data source categories via "View list", or from the side menu. It presents the operating status of every registered OPC server connection in a 12-column table.
| Column | Width | Display |
|---|---|---|
| No. | 60px | Row number |
| Connection ID | 100px | OPC domain badge |
| OPC Name | 250px | Display name |
| Status | 70px | Full-color gradient status dot (normal/delayed/abnormal/disconnected/unknown) |
| Protocol Type | 60px | Dark badge — protocol-badge-opc-ua (navy) / protocol-badge-opc-da (purple). Black background with white text, instantly distinguishable from badges in other columns |
| Program ID | IP | Auto | ProgID or IP address of the OPC server |
| Start | 60px | ▶ Start collection / ◼ Stop (shown according to current state) |
| Collection Start Time | 120px | Time collection began (scan_start_time) |
| Connected Tags | 70px | Number of tags mapped to that OPC |
| Total / Received Points | 80px | Cumulative / today's received point count |
| Errors | 60px | Cumulative error count (highlighted in red) |
| Actions | 80px | 🔧 Edit / 🗑 Delete (active only for roles with change permission) |
Use the ↻ icon button to the right of the table header to refresh. A help box at the bottom of the screen provides OPC connection recommendations (such as keeping each connection under 1,000 tags).
Adding a New Connection
Clicking the Connect link on a card opens the registration form for that data source type. The input fields differ by data source, but the following are common to all.
| Common input | Description |
|---|---|
| Site | Select the site where the connection will be registered |
| OPC ID or name | Domain identifier (letters/digits/_ only recommended) |
| Connection information | Host/port, authentication, topic, file path, etc. (varies by data source) |
| Collection cycle and options | Polling interval, start/end time, etc. |
| Description | Notes |
After registration, and until you confirm actual data ingestion on the Connection Status screen, only the metadata is registered. Once collection begins, a "Connected" badge appears on the card in Connection Status.
Status Screen (/connect/status)
Monitor the status of all live data source connections on a single screen.
Top Toolbar
📡 연결 상태 [총 N 개의 데이터소스 연결] [사이트 ▼] [검색] [↻]
| Control | Width | Description |
|---|---|---|
Site selector (site_id) | 200px | All sites + list of registered sites (사이트명 [설명]) |
| Search | Red + 🔍 | Applies the site filter and refreshes the cards |
| Refresh | Red + ↻ | Refreshes the screen |
Left — Statistics Charts and History
The left col-lg-2 area stacks five charts and three history panels vertically.
1) Connection Status Distribution (connection_stat_chart, 200px)
Shows the distribution of the five states (Connected, Connection Delayed, Connection Abnormal, Disconnected, Not Observed) across all registered connections as a donut chart.
2) Communication Connection Status Trend [Today] (connection_trend_chart, 150px)
Trend of connection counts by state per time slot (last 24 hours).
| Color | Meaning | Threshold |
|---|---|---|
| 🟢 Green | Connected | Data received within 10 minutes |
| 🟠 Orange | Connection Delayed | No data for 10 minutes or more |
| 🔴 Red | Connection Abnormal | No data for 1 hour or more |
| 🟥 Dark red | Disconnected | No data for 24 hours or more |
| ⬜ Gray | Not Observed | No data reception history |
ℹ️ Hovering over the information icon in the header displays the five-level criteria above as a tooltip.
3) Communication Latency Status Trend [Today] (latency_trend_chart, 150px)
| Color | Meaning | Threshold (maximum latency in interval) |
|---|---|---|
| 🟢 Green | Normal | Under 10 seconds |
| 🟠 Orange | Delayed | 10 to 60 seconds |
| 🔴 Red | Severely delayed | 60 seconds or more |
The tooltip reads "average latency under 1 second / 1~10 seconds / 10 seconds or more", but the
calculation that builds the bars uses the maximum latency in that time interval (max_latency), with thresholds of 10 and 60 seconds. When reading the bars, use
the table above.
That is why a connection with 3 seconds of average latency appears green on this chart. The average-based 1-second/10-second thresholds are a separate criterion used to determine asset health status — Status Code Definitions lays out both criteria side by side.
4) Point Validation Failure Trend [Today] (point_fail_trend_chart, 150px)
Aggregates, by hour, the number of received data points that failed type, range, or validity checks.
5) Recent Disconnections [Last 24 Hours] (connection_status_history, 200px scrollable)
History of data sources disconnected during the last 24 hours. Listed in reverse chronological order — data source ID, time of disconnection, last normal time.
6) Recent Reception Delays [Last 24 Hours] (connection_latency_history, 200px scrollable)
History of data sources that experienced response delays during the last 24 hours.
7) Recent Point Validation Failures [Last 24 Hours] (point_validation_status_history, 400px scrollable)
History of points that failed validation during the last 24 hours.
Right — Data Source Card Grid
The right col-lg-10 area displays data source cards in a grid, grouped by category.
Card Categories (from the top)
| Section header | Display |
|---|---|
Edge Gateway / File / API [N] | FILE-type data sources (Edge GW, files, APIs, etc.) |
PLC [N] | PLC data sources |
| (followed by OPC, Modbus, external DB, etc.) | By registered data source type |
Card Composition (opc_obj)
┌─────────────────────┐
│ [SUB_TYPE] │ ← 작은 글씨
│ OPC_ID │ ← ID
│ ──────────────── │
│ 현재 상태값(_mm) │ ← 매분 갱신 상태 표시
└─────────────────────┘
| Area | Display |
|---|---|
| Top | Data source subtype (e.g., FILE, MQTT, REST_API) |
| Middle | Data source ID — ellipsis (...) when truncated |
| Bottom | Per-minute status display (${opc_id}_mm) — colors map to the five states |
Clicking a card opens the detail screen for that data source (
/connect/view?opc_id=...), where you can review collection history, the tag list, and logs. Hovering shows the card's description as a tooltip.
Empty State Message
When no data source is registered in that category:
No connected agents
Data Source Detail Screen (/connect/view)
Reached by clicking a data source card on the Status screen. This view gathers all operating information for a single OPC/PLC/DB connection on one screen.
Page Header
📶 프로토콜 연결 보기 [📅 날짜 선택 ▼] [날짜 이동 →] [← 목록] [↻]
{site_name} / {opc_name}
| Control | Action |
|---|---|
| Date picker | Calendar — query connection status for another date |
| Go to date → | Reload with the selected date |
| ← List | Return to the Status screen |
| ↻ | Refresh the screen |
Left ① Basic Connection Information + Expand
Two buttons on the right of the panel header:
| Button | Action |
|---|---|
⌄ Expand (viewOPCDetail()) | Toggles display of the site information section |
| ⚙ Connection settings | Goes to the edit form for that data source (change permission required) |
Basic Information (ds-dl)
| Item | Display |
|---|---|
| Connection ID | OPC domain badge |
| Connection name | Data source display name |
| Description | Notes |
| Registration date | Registration timestamp (monospace) |
Protocol Connection Information (ds-dl)
| Item | Display |
|---|---|
| Connection type | OPC_UA / OPC_DA / PLC / MODBUS / DB / FILE, etc. |
| Subtype | Manufacturer or detailed protocol (e.g., SIEMENS_S7, MITSUBISHI_MC) |
| Agent IP / port | IP and port of the agent (Edge) (monospace) |
Site Information (only when expanded)
| Item | Display |
|---|---|
| Site ID | Site domain badge |
| Site name | Site display name |
Left ② Connection Statistics (Latency and 5 States)
Today's Peak Latency Trend (3 cells)
| Cell | Display |
|---|---|
| Minimum | Fastest response latency (ms, comma-separated) |
| Average | Average response latency |
| Maximum | Slowest response latency |
Below is a 24-hour latency trend line chart (latency_24hh) — if you see an unusual spike, cross-check that time on the Status screen.
Connection Abnormal Count
A read-only field showing today's cumulative "Connection Abnormal" count.
5-Cell Status Statistics Grid
┌─🟢 연결─┐ ┌─🟠 지연─┐ ┌─🔴 이상─┐ ┌─⬛ 끊김─┐ ┌─⚪ 미관측┐
│ 12,453│ │ 23 │ │ 5 │ │ 1 │ │ 0 │
└────────┘ └────────┘ └────────┘ └────────┘ └────────┘
Each of the five colored cells shows its cumulative count as a large number, with a small donut chart (connection_priority_count_chart) displayed underneath.
Right ① Connection Status Timeline + LIVE Dot
Panel header subtext: "Today's connection/disconnection timeline [{status_txt}]". An ⓘ information icon on the right of the header shows a guidance tooltip on hover.
Left — Current Status Card (live)
A status card with a large icon and label (last_connection_status).
| Icon + label | Meaning | Dot toggle |
|---|---|---|
❓ Not Observed (is-mute) | No data reception history | ⚪ Gray |
🟢 Connected (is-ok) | Data received within 10 minutes | 🟢 LIVE dot (pulsing) |
🟠 Connection Delayed (is-warn) | No data for 10 minutes to 1 hour | 🟠 |
🔴 Connection Abnormal (is-err) | No data for 1 to 24 hours | 🔴 |
⬛ Disconnected (is-disc) | No data for 24 hours or more | ⬛ |
When the 🟢 LIVE dot pulses, data is arriving in real time. Hovering shows the last reception time as a tooltip.
Right — 24-Hour Timeline Bar (opc_connection_timeline)
Displays 24 hours as a horizontal bar, with each time slot colored according to one of the five connection states. Hovering over the bar shows the state at that time as a tooltip.
Right ② 24-Hour Point Reception Chart
Panel header subtext: "Today's per-minute reception count over 24 hours".
A line chart of points received per minute (opc_point_trend_24h). A sharp drop compared with normal business hours means data ingestion stopped at that time. While loading, a "Loading..." message appears in the center.
Right ③ Connected Tag List
A list of real-time values for the tags mapped to this data source (7 columns).
| Column | Width | Display |
|---|---|---|
| Tag ID | 80px | Domain badge |
| Tag name | 350px | Tag display name |
| Data type | 100px | INT/FLOAT/STRING/BOOL, etc. |
| Current value | Auto | Last received value (quality color badge) |
| Last update time | Auto | Reception timestamp |
| Latency [MS] | Auto | Average response latency |
| Actions | 50px | 🔍 → go to that tag's detail |
The table updates in real time over WebSocket. If a value stops flickering, reception for that tag has stopped.
Edge Screen (/edge/index)
Monitor the system status and PLC collection statistics of edge devices installed in the field on a single screen. The page subtitle reads "Monitors edge connection status and PLC data collection in real time."
Top Toolbar
📡 엣지 [사이트 ▼] [검색] [↻]
| Control | Width | Description |
|---|---|---|
Site selector (site_id) | 200px | All sites + registered sites |
| Search | 🔍 | Applies the site filter |
| Refresh | ↻ | Re-queries the list immediately |
The list is re-queried automatically every 60 seconds. The "every 10 seconds" mentioned in the help box is not the screen refresh cycle but the interval at which the edge reports its own status to the server — the two are easy to confuse. To see a value that just changed, press ↻.
List Table (8 Columns)
Related metrics are grouped into a single column. CPU, memory, and disk no longer occupy one column each as they did previously.
┌────────────────────────── 목록 ──────────────────────────┐
│ 연결 │ EDGE_ID │ 엣지 │ 리소스 │ 네트워크 │ PLC │ 처리량 │ 액션 │
└──────────────────────────────────────────────────────────┘
| Column | Width | Display |
|---|---|---|
| Connection | 60px | Connection status icon badge (🔌 Disconnected / ⚠ Warning) |
| EDGE_ID | 130px | Edge device identifier |
| Edge | 260px | Edge name (+ O/S hostname and version) |
| Resources | 150px | CPU, memory, and disk usage in one cell |
| Network | 110px | NET 1 / NET 2 / WiFi / VPN status in one cell |
| PLC | 80px | Number of PLC connections normal / disconnected |
| Throughput | 110px | Current collection rate (mps) · Σ cumulative received points + hourly reception sparkline for the last 24 hours |
| Actions | 110px | View details, etc. |
In the throughput sparkline, an empty interval means "not observed", not "0 received". Hovering over the cell shows the same explanation as a tooltip.
Edge Help Box
The following operating guidance (four items) is shown at the bottom of the page.
❓ Edge Help The Edge menu lets you check edge connection status and PLC data collection status.
- For any edge shown as Disconnected [🔌], be sure to check the physical environment, the network, and the edge O/S.
- An edge shown as Warning [⚠] has S/W errors during collection; connect to the edge and resolve the problem.
- When CPU/memory/disk usage reaches 80% or more, check the edge and free up resources.
- Edge status is updated every 10 seconds.
Operating Procedure (When an Anomaly Is Found)
- Edge with 🔌 (Disconnected) connection → field inspection (power, network, hardware)
- CPU/memory/disk at 80% or more → connect to the edge and clean up unnecessary processes and logs
- Non-zero PLC disconnection count → inspect that edge's PLC cabling and configuration
Edge Detail Screen (/edge/view)
The detail screen for a single edge, opened by clicking the EDGE_ID or the "Actions" column in the edge list. From this one screen you can:
- Check registration metadata, host, network, and security keys
- View real-time CPU, memory, and disk gauges
- View eight PLC collection statistics
- View the OPC servers connected to the edge and start/stop collection
- View the Docker containers running on the edge and start, stop, restart, log, or delete them
- Deploy a new container remotely
- Restart the edge application and reboot the edge host OS
⚠️ Every control command on this screen executes directly on the field edge device. On an edge in production, stopping OPC collection, deleting a container, or rebooting the system interrupts data collection. Always perform these actions during a maintenance window.
Page Header and Refresh
🛜 ${엣지명} [← 목록] [⚡ 실시간 새로고침] [↻ 새로고침]
사이트: ${site_id} · 엣지 ID: ${edge_id} · ${설명}
| Control | Action |
|---|---|
| ← List | Return to /edge/index |
| ⚡ Real-time refresh | Calls the edge device directly to update the gauges, OPC list, and Docker list with current values |
| ↻ Refresh | Refreshes the entire page |
Only two things run automatically on this screen.
| Target | Cycle | Source |
|---|---|---|
| Stored values such as the left information panel and PLC statistics | 30 seconds | Read-model stored on the Platform (/edge/{id}/model) — values the edge reports via a roughly 60-second heartbeat |
| Upgrade availability check | 5 minutes | — |
The system resource gauges, the OPC list, and the Docker container list call the edge directly only once, when you open the screen. After that, they are re-read only when you press ⚡ Real-time refresh. This is deliberate, to avoid hammering the device every second.
So if you leave the screen open and watch the gauges, those values date from the moment you opened it. To verify the result of something you did in the field, always press ⚡.
Connection / Health Timeline Tabs
The statistics area is split into two tabs.
| Tab | Content |
|---|---|
| Connection Status Timeline (selected by default) | Reachability over the last 24 hours |
| Health Status Timeline | PLC connection and collection over the last 24 hours |
The health timeline is drawn when you click the tab and it becomes visible. It is normal for it to be empty before you open the tab.
Left — Four Panels: Basic Info, Host, Network, Security
Four panels are stacked vertically in the left col-lg-3 area. All use a definition list (label / value) format.
① Basic Information — Includes the System Reboot Button
| Item | Meaning |
|---|---|
| Edge ID | Domain ID badge (EDGE domain color) |
| Name | edge_name |
| Site | Site ID (domain badge) |
| Role | role (- if none) |
| Status | status value from the master table |
| Description | Notes |
| Registration date | Date and time the edge was first registered |
| Last status | Date and time the last heartbeat was received |
On the right of the panel header is a 🔴 red ⏻ power icon button — clicking it reboots the edge host OS (see Safe Procedures below).
② Host
| Item | Meaning |
|---|---|
| Hostname | Operating system hostname |
| Version | Edge software build version |
| Build date | Date built |
| CPU cores | Number of cores |
| Threads | Number of threads |
| Temperature | CPU temperature (°C) — 0 if the board has no temperature sensor |
③ Network
IPs for six interfaces: VPN / NET 1 / NET 2 / WiFi / management network / GSM. Shows - when unconfigured.
| Item | Purpose |
|---|---|
| VPN IP | Edge-side address of the headquarters ↔ field VPN tunnel |
| NET 1 | Primary LAN |
| NET 2 | Secondary LAN |
| WiFi | Wireless |
| Management network | OOB management network |
| GSM | LTE/3G backup line |
④ Security
| Item | Display |
|---|---|
| API Key | The key the edge uses to authenticate to the server — masked as first 4 characters + **** + last 4 characters |
This is the same key as the X-API-Key on the Flow REST node — this screen shows only the masked form; the actual key value is never displayed.
Right ① System Resource Gauges + App Restart
The panel header shows the ⓘ label "CPU · Memory · Disk usage (real time)", with an App Restart button on the right.
Status Line (state-row)
A single-line status message appears above the gauges.
| Display | Meaning |
|---|---|
🟢 엣지 LIVE — 정상 응답 중 | The edge responded to ping |
🔴 엣지 OFFLINE | The edge's /api/v1/monitoring is ping=false |
🔴 엣지 UNREACHABLE | The HTTP request itself failed (network outage) |
Three Gauges (140px donut each)
| Gauge | Unit | Color thresholds |
|---|---|---|
| CPU usage | % | < 80 green · 80~89 orange · ≥ 90 red |
| Memory usage | % | Same |
| Disk usage | % | Same |
Values are shown to one decimal place (e.g., 73.4%).
App Restart Button
Restarts only the edge application (daemon). The host OS is unaffected.
🔄 Pressing it opens a confirmation dialog → "Restart" → after about 3 seconds the edge process restarts itself and returns after about 10 seconds. During that time, the gauges and statistics temporarily show
-.
Right ② PLC / Collection Statistics (8-Cell stat-box)
Eight statistics boxes are laid out in a grid.
| Box | Unit | Meaning | Emphasis |
|---|---|---|---|
| Connections normal | count | Number of live PLC connections | 🟢 ok |
| Connections disconnected | count | Number of dropped PLC connections | 🔴 err |
| Read successes | count | Cumulative counter of successful PLC polls | - |
| Read failures | count | Cumulative counter of failed PLC polls | 🔴 err |
| Points sent | count | Cumulative number of points sent to the server | - |
| MPS | count/sec | Message processing rate (points per second) | - |
| Average latency | MS | Average response latency in the recent window | - |
| Peak latency | MS | Maximum response latency in the recent window | - |
The counters (read successes, read failures, points sent) restart from 0 whenever the edge process restarts. Interpret them as a trend, not as absolute values.
Right ③ OPC Status + 5 Summary Chips + Start/Stop Collection
The panel header reads "OPC servers connected to this edge and their collection status".
Five Summary Chips (opc-chip-v2)
Each chip consists of an icon, a large number, and a label.
| Chip | Icon | Meaning | Color |
|---|---|---|---|
| Registered | ☰ | Total number of OPC servers mapped to this edge | Gray |
| Connected | ✓ | Number of connection_status = CONNECTED | 🟢 Green |
| Collecting | ≋ | Connection normal + scan_status = START | 🟢 Green |
| Stopped | ⏸ | Number of scan_status = STOP | Gray |
| Disconnected | 🔗∕ | Number of connection_status ≠ CONNECTED | 🔴 Red |
List Table (7 Columns)
| Column | Width | Display |
|---|---|---|
| OPC ID | 160px | 🔌 + ID (click to open the /connect/opc/edit/{id} edit screen) |
| Name | Auto | opc_name |
| Type | 80px | OPC UA / OPC DA badge |
| Agent IP | 140px | OPC agent IP (monospace) |
| Connection | 90px | 🟢 CONNECTED / 🔴 otherwise — shown as a dot |
| Collection | 90px | 🟢 START / 🟡 otherwise / ⚫ STOP — dot |
| Control | 160px | Action buttons (below) |
Control Buttons
| State condition | Button shown | Action |
|---|---|---|
scan_status ≠ START | 🟢 ▶ Start collection | Confirm → POST /edge/{edge_id}/opc/{opc_id}/start → screen reloads after 2 seconds |
scan_status = START | 🔴 ◼ Stop collection | Confirm → POST /edge/{edge_id}/opc/{opc_id}/stop → screen reloads after 2 seconds |
| Always shown | ⚙ Edit OPC | Go to the /connect/opc/edit/{opc_id} edit screen |
The stop-collection dialog is shown in ⚠️ warning color (orange) — data really will stop flowing.
Right ④ Docker Containers (Remote Deploy, Start, Stop, Logs, Delete)
Lists the containers running in Docker on the edge host, refreshed every 5 seconds.
Panel Header
| Button | Action |
|---|---|
| ↻ | Refresh immediately |
| ➕ New deployment | Container deployment dialog (name, image, ports, environment variables) |
List Table (6 Columns)
| Column | Width | Display |
|---|---|---|
| Name | 160px | ▢ cube icon + container name (bold) |
| Image | Auto | Image tag (monospace, e.g., grafana/grafana:latest) |
| Status | 80px | 🟢 UP / 🔴 DOWN chip |
| Last status | 200px | docker status text (e.g., Up 3 hours) |
| Ports | 200px | 호스트:컨테이너/프로토콜, comma-separated |
| Control | 240px | Five action buttons (below) |
Control Buttons (Same for Every Row)
| Button | Color | Action |
|---|---|---|
| ▶ Start | 🟢 Green | POST /app/{name}/start |
| 🔄 Restart | 🟡 Orange | POST /app/{name}/restart |
| ◼ Stop | 🔴 Red | POST /app/{name}/stop (warning dialog) |
| 📄 Logs | Gray | GET /app/{name}/logs → shows the container's standard output in a wide dialog |
| 🗑 Delete | Red text | DELETE /app/{name} (danger dialog) — forces termination even if running |
New Deployment Dialog
| Input | Example |
|---|---|
| Container name | grafana |
| Image | grafana/grafana:latest |
| Ports (host:container, comma-separated) | 3000:3000 |
| Environment variables (KEY=VAL, comma-separated) | GF_SECURITY_ADMIN_PASSWORD=admin |
The Docker daemon on the edge host receives the deployment and performs the image pull and run. Large images take time, so refresh the list to check progress.
Safe Procedures for System Reboot and App Restart
🔄 App Restart (Edge Application Only)
- Click App Restart at the top right of the system resource panel
- Confirmation dialog → "Restart"
- The process restarts itself after about 3 seconds
- It returns after about 10 seconds — gauges and statistics repopulate automatically
Impact: PLC/OPC collection is interrupted for about 10 to 15 seconds. CEP alarm evaluation continues on the server side. Cumulative counters restart from 0.
⏻ System Reboot (Edge Host OS)
The 🔴 ⏻ button in the header of the left ① Basic Information panel.
- The confirmation dialog is shown in danger (red)
- There is no response for about 1 minute; the edge returns only after the OS has fully booted
- Both PLC and OPC are disconnected during boot — containers also stop and restart
Use this only in the following cases: ① memory/disk is full at 100% and an app restart does not recover it; ② the board itself has become slow to respond. In general, try App Restart first.
Usage Scenarios
| Scenario | Screen | Procedure |
|---|---|---|
| Register OPC for a new line | Protocol → OPC card → Connect | Enter site, OPC ID, server address → Save |
| Bulk check of disconnections in production | Status | Left "Recent Disconnections" 24-hour history → identify red and dark-red cards on the right |
| Analyze causes of point validation failures | Status | Left "Point Validation Failure Trend" + "Recent Validation Failures" → asset ID + timestamp → compare on the trend screen |
| Inspect edge device resources | Edge | Identify devices at 80% or more CPU/memory/disk → clean up in the field |
| Identify sites with PLC disconnections | Edge | Sort by the "PLC Disconnected" column → inspect edges at sites with non-zero values |
| Analyze network latency | Edge | Sort by "Peak Latency [MS]" → check the network on devices at 10 seconds or more |
| Restart Grafana or integration containers on an edge | Edge detail → Docker | Click 🔄 Restart on the container row (no full page refresh needed) |
| Pause a field OPC for inspection | Edge detail → OPC | ◼ Stop collection on that OPC row → after inspection, ▶ Start collection |
| Recover a slow-responding edge | Edge detail → System resources | App Restart (10-second impact). If that does not recover it, ⏻ System Reboot (1-minute impact) |
| Deploy a new monitoring tool remotely | Edge detail → Docker → ➕ New deployment | Enter image, ports, and environment variables and deploy — the image pull is automatic |
Frequently Asked Questions
Q. I registered an OPC data source, but no card appears on the [Status] screen. A. Immediately after registration, the card should appear in the "Not Observed" (gray) state. If it does not appear, check that you selected the site where that OPC is registered in the site selector.
Q. The trend charts on [Status] are empty. A. Either there is no data source at all for that site, or there has been no ingestion history today. If you just registered the data source, wait a moment for the first ingestion.
Q. The tooltip on the ℹ information icon above a chart appears cut off. A. Hovering over the information icon displays a table of the five-level status criteria. On a narrow screen it may be clipped, so widen the window or refer to the Status Code Definitions screen.
Q. The edge status is not updating. A. Three overlapping cycles make this easy to confuse: ① the edge reports its status to the server every 10 seconds, ② the list screen re-queries the server every 60 seconds, and ③ the detail screen refreshes its stored values every 30 seconds. So a value in the list lagging by up to about a minute is normal. If it is frozen longer than that, the edge is unable to respond (disconnected). Perform a field inspection or check the EDGE application logs under Diagnostics.
Q. The CPU gauge on the detail screen keeps showing the same value. A. The gauges, OPC list, and Docker list do not refresh automatically — they call the edge directly only once, when you open the screen. Press ⚡ Real-time refresh.
Q. The PLC disconnected column went up to 1. How do I investigate? A. Use the action button for that edge to open the detail screen, where you can see per-PLC status. Alternatively, identify red cards in the PLC card grid on the Status screen.
Q. One of my edges is at 100% CPU/memory. A. Connect to the field immediately and clean up unnecessary processes. If it recurs, consider upgrading the device specification or adjusting the collection cycle.
Q. I clicked a card and got a blank screen.
A. You are taken to the data source detail screen (/connect/view), but with no incoming data it can look blank. Rich information is shown only for data sources with an ingestion history.
Q. Does it refresh automatically in real time?
A. The per-minute card display (_mm) on the [Status] screen refreshes automatically. The statistics charts on the left refresh at a fixed interval; to apply changes immediately, press the ↻ refresh button at the top right.
Q. The "received point count" suddenly drops to 0. A. When the edge counter restarts, it starts again from 0. To tell a genuine 0 from a counter reset, also check the per-site data counts on the Daily Statistics screen.
Q. I stopped OPC collection on the edge detail screen (/edge/view), but after a while it goes back to START.
A. The OPC itself may have the auto-restart option enabled, or an automatic start command may have arrived from a workflow. Check the auto-restart setting on the OPC Edit screen. If you need a permanent stop, disable the OPC itself.
Q. The Docker container panel is empty. A. ① The edge may not be in Docker mode — some edges start directly under systemd. ② The Docker daemon on the edge OS may not be responding (in which case ↻ Refresh still shows nothing). Consider a field inspection or a ⏻ system reboot.
Q. The container ▶ Start, 🔄 Restart, and ◼ Stop buttons are all visible. Which one is active? A. All buttons are always clickable — the command is sent regardless of Docker state. Pressing ▶ Start on an already-running container may be silently ignored or may restart it, so check the status chip (🟢 UP / 🔴 DOWN) and choose the command you intend.
Q. What is the difference between ⏻ System Reboot and 🔄 App Restart? A. ① App Restart (10-second impact) — restarts only the edge daemon process; OPC/PLC collection is interrupted for 10 seconds. ② System Reboot (1-minute impact) — boots the entire edge host OS, restarting Docker containers as well. In general, always try App Restart first.
Q. The API Key is masked. How do I see the actual value? A. The masking is intentional — the plaintext key is not shown anywhere in the UI. It is the same key as the X-API-Key on the Flow REST node, so store it somewhere safe when it is first issued. If you lose it, ask your administrator to reissue it.
Related Screens
- Dashboard — site map and connection rate on the KPI cards
- Daily Statistics — comparative analysis by site, OPC, and tag
- Data Points — review ingested data as charts
- Factory Management — register the site and asset tree
- Diagnostics — diagnostic logs for edge and server modules
- Status Code Definitions — the five connection states and their color mapping