Real-Time Functions, Executed on Your Nodes

The Live tab gives you real-time access to Netdata Functions that can be executed on any node running the Netdata Agent. These specialized routines, provided by various collectors, deliver live information from your monitored nodes - including process explorers, database monitoring functions, network topology maps, and more. You can use these Functions to retrieve extra diagnostic information and perform troubleshooting actions directly on the node.

The Live tab provides access to several built-in Functions: Process Explorer (replaces top/htop with browser-based process monitoring), Top Queries (database query performance for PostgreSQL, MySQL, SQL Server, Oracle, ClickHouse, MongoDB), Network Connections (active connections ranked by bandwidth), Systemd Services (service resource usage), Block Devices (disk I/O like iostat), and Mount Points (filesystem usage like df). Additional Functions are provided by individual collectors.

The Metrics tab shows time-series data collected and stored by the agent - historical charts you can scroll through. The Live tab executes Functions on your nodes in real-time to retrieve current diagnostic information that isn’t stored as time-series. Think of it as running console commands (top, netstat, iostat) from your browser, with the added benefits of ML context, filtering, and team collaboration.

Netdata tracks both running and exited processes, accumulating metrics from entire process families. This captures resource usage from short-lived subprocesses that traditional monitoring misses - critical for shell scripts, CI/CD pipelines, and automation jobs. Plus, PSS (Proportional Set Size) memory accounting reveals true memory consumers versus processes merely sharing libraries.

Yes. Deploy Netdata’s DaemonSet and every pod, container, and process appears automatically with full Kubernetes label integration (namespace, pod, deployment). Per-container CPU, memory, I/O, and network metrics reveal resource limit utilization, throttling, and OOM risks - with zero configuration or instrumentation required.

During incidents, highlight any timeframe on any chart (15 seconds minimum). Anomaly Advisor analyzes ALL metrics across ALL nodes simultaneously, ranking them by ML-powered relevance. The root cause typically surfaces in the top 30-50 results. You can then use Live tab Functions to dig deeper - examining specific processes, database queries, or network connections on the affected node.

The Process Explorer provides browser-based access with the same per-second precision as console tools, plus historical data, ML anomaly detection, and team collaboration. No SSH access required, works through firewalls, and shows what happened during incidents - not just current state. It’s console-quality debugging without infrastructure access.

Most resource issues last seconds, not minutes. A process spiking to 100% CPU for 5 seconds every 30 seconds appears as harmless 17% average in minute-level monitoring. Netdata’s per-second data captures microbursts, transient spikes, and brief resource hogs that averaging hides - revealing the truth, not statistical approximations.

Yes. Netdata automatically correlates from Kubernetes cluster to namespace to pod to container to process to PID. Click any resource spike and see the complete path: which pod, which container, which process category, which specific PID. No context switching between tools, no manual timestamp correlation - just instant visibility.

Yes. The Top Queries function retrieves query performance data directly from your database engines. PostgreSQL uses pg_stat_statements, MySQL uses performance_schema, SQL Server uses Extended Events, Oracle uses V$SQLSTATS, ClickHouse uses system.query_log, and MongoDB uses system.profile. Sort by CPU time, elapsed time, logical reads, physical reads, executions, or rows returned. See Database Monitoring for details.

Yes. Netdata provides real-time deadlock detection for SQL Server (deadlock graphs via Extended Events), PostgreSQL (via pg_stat_activity and pg_stat_database), and MySQL (InnoDB deadlock counts). Alerts fire immediately when deadlocks occur, not after users complain.

Install Netdata (60 seconds), and all Functions become available automatically. The Process Explorer, Network Connections, Systemd Services, and other Functions work immediately. Database Top Queries require the corresponding collector to be configured. No dashboard building, no query language learning - just instant access.

Yes. Netdata complements existing tools through OpenTelemetry support, Prometheus export, and Grafana integration. Use the Live tab for real-time diagnostics and detailed troubleshooting while maintaining your existing observability stack. Visit our integrations page for the complete list.

None. The Live tab is designed for operations teams of all skill levels. Junior engineers get senior-level insights through intuitive interfaces that require zero query language knowledge. Automatic discovery, interactive sorting and filtering, and ML-powered prioritization make every team member effective from day one.

Netdata tracks resources from creation to deletion, capturing complete lifecycle metrics even for containers lasting seconds. Accumulated metrics from exited children reveal total resource impact of ephemeral workloads. Historical data shows what happened during incidents, not just current state - critical for dynamic Kubernetes environments.