Operations Monitoring

The Operations module consolidates the platform runtime metrics, system information, logs, background tasks, export records, and WebRTC session status for routine inspection and fault isolation. The page and its data APIs require the Administrator role.

Prerequisites

Item	Content
Menu entry	Manage → Operations
Route	/monitor, enters /monitor/overview by default
Page permission	"View Ops Monitor"
Visible roles	Administrator
Data APIs	The monitoring, queue, and log APIs validate the Administrator role
Dependencies	Redis (Operations metrics stream and task queues), database

Subpages

Subpage	Route	Content shown	Data source
Monitoring Overview	/monitor/overview	Current values and trends of database latency, Redis latency, and queue backlog	Redis Operations metrics stream
System Information	/monitor/status	Version, build date and time, Node.js version, Uptime, Hostname, Operating System, Architecture, CPU cores, Total Memory	Server build information and host information
System Log	/monitor/logs	Access, active user, login, operation, and workflow logs	Access and audit data
Background Tasks	/monitor/jobs	Task list and actions of the system queue and export queue	Task queues
Export Records	/monitor/exports	Export task history and artifacts	Export records
WebRTC	/monitor/webrtc	Active rooms, publishers, viewers, and track count	WebRTC service room list

Runtime Metrics

Monitoring Overview

The top of /monitor/overview shows three metric cards, with the corresponding trend charts below.

Metric	Meaning	Threshold
Database latency	Database probe duration	Below 50 milliseconds is normal, 50–200 milliseconds is a warning, and 200 milliseconds or more is critical
Redis latency	Redis PING duration	Below 10 milliseconds is normal, 10–50 milliseconds is a warning, and 50 milliseconds or more is critical
Queue backlog	Total number of waiting tasks across queues	Above 100 is a warning, and above 1000 is critical

Time range	Purpose
1 hour	Real-time isolation
24 hours	Routine inspection
7 days	Trend analysis

The overview summary refreshes automatically every 30 seconds; the "Collect Now" button triggers one manual metric collection and refreshes the summary and trends.

System Information

/monitor/status lists application and host information.

Field	Description
Version, build date, build time	Application build information
Node.js version	Runtime version
Uptime	Time the process has been running
Hostname, Operating System, Architecture	Host information
CPU cores, Total Memory	Host resources

Background Tasks

/monitor/jobs manages the tasks of two queues.

Queue	Name	Tasks handled
System queue	iojobs	System-level tasks such as data preprocessing
Export queue	ioexports	Data export tasks

Task status	Description
waiting	Waiting
active	Running
completed	Completed
failed	Failed
delayed	Delayed
paused	Paused

Action	Effect	Constraint
Pause Queue	Pauses processing of the queue	Affects the execution of new tasks
Resume Queue	Resumes processing of the queue	None
Clear Waiting	Deletes all waiting tasks of the queue	Cannot be undone
Clean History	Deletes completed and failed tasks older than 24 hours	None
Retry Failed	Retries failed tasks in bulk	Up to 1000 per operation
Retry	Re-runs a single failed task	Failed tasks only
Stop	Stops a running task	Running tasks only
Remove	Deletes a waiting task	Waiting tasks only

warning

Pausing a queue affects the execution of new tasks, and clearing waiting tasks deletes the waiting tasks.

WebRTC Monitoring

/monitor/webrtc lists the currently active rooms.

Column	Description
Agent ID	Device Agent ID and publisher IP of the room
Status	Online, Disconnect
Start time	Session start time
Track count	Number of media tracks in the room
Viewers	Number of viewers and their IPs
Actions	Force disconnect

Force disconnect requires confirmation and a reconnect-ban duration in hours; a value of 0 means no ban.

Record Queries

System Log

Subpage	Route	Content	Filter fields
Log overview	/monitor/logs	Today's failed logins, active users, key operations, and data browsing; recent failed logins and locked users	None
Access Log	/monitor/logs/visit	Data, account, viewing IP, viewing time	Data, account, IP, time
Current active users	/monitor/logs/active	Username, active time, IP address	Time range 1 day, 7 days, 30 days
Login Logs	/monitor/logs/login	Username, result, IP address, login type, error message, time	User, username, IP, time range, result
Operation Logs	/monitor/logs/operation	Username, operation content, IP address, time	User, IP, time range
Workflow logs	/monitor/logs/workflow	Rule ID, time, description	Workflow, dataset, match rule, action rule, time range

The Log overview provides an immediate unlock entry for locked users. On the Current active users page, an Administrator disables an account. Module permission changes are written to a separate audit record, and the Operations module provides no page to view that record.

Export Records

/monitor/exports requires the "View Export List" permission, and re-exporting requires the "Create Export Job" permission.

Field	Description
ID, filename, type	Export task identifier and artifact
Export parameters, status, user, dataset, creation time, completion time	Task information
Status	Queued, Processing, Running, Completed, Failed
Type filter	lerobot, hdf5, mcap
Actions	Re-export, which requeues the task; completed records allow downloading the artifact

Troubleshooting

Symptom	Possible cause	Action	Owner
No data in the overview	Metrics not yet collected, or Redis is unavailable	Click "Collect Now"; check the Redis connection	Operations
Database or Redis metric marked critical	The corresponding service responds slowly, the connection pool is exhausted, or the network jitters	Check the load and connections of the corresponding service	Operations
Queue backlog keeps growing	Task volume exceeds capacity, the Worker is not running, or a single task takes too long	Check the Worker status and, if necessary, pause the queue and clear waiting tasks	Operations
A task stays running for a long time	Worker failure, or the task is blocked	Stop the task and retry	Operations
No data on a log page	No records in the time range, or the filter is too narrow	Widen the time range or clear the filter	Administrator
WebRTC room cannot be force-disconnected	The request failed	Retry the operation and check the WebRTC service status	Operations

Related Pages

Page	Purpose
System Settings	System parameters and log-related configuration
Dataset Export	Creation and quota of export tasks
Module Permissions	Configuration of module permissions
Data Management	Locating a specific dataset
Trash	Restoring deleted data