TUI Guide

The bewitch TUI provides 8 views for real-time system monitoring with historical charts.

launch

bewitch

# or connect to a remote daemon
bewitch -addr myserver:9119 -token my-secret

Views

Views are accessed via number keys. Tab numbering is fixed regardless of hardware availability.

Key	View	Description
`1`	Dashboard	Multi-column overview of all subsystems
`2`	CPU	Per-core usage with historical chart
`3`	Memory	RAM and swap breakdown with history
`4`	Disk	Per-mount space, I/O rates, SMART health
`5`	Network	Per-interface throughput with bits/bytes toggle
`6`	Hardware	Temperature, power (RAPL), ECC memory, and GPU sub-sections
`7`	Process	All processes, sortable, searchable, pinnable
`8`	Alerts	Alert rules and fired alerts

Navigation

Key	Action
`Tab` / `Shift+Tab`	Cycle views forward / backward
`←` / `→` or `h` / `l`	Cycle views forward / backward
`<` / `>`	Cycle history time range (1h, 6h, 24h, 7d, 30d)
`q`	Quit

Network View

Per-interface throughput with sparklines and historical chart. Select which interfaces appear on the chart.

Key	Action
`j` / `k`	Navigate interface list
`Space`	Toggle interface in chart
`a`	Select / deselect all
`b`	Toggle bits/bytes display

Hardware View

The Hardware view combines temperature sensors, power consumption (RAPL), ECC memory errors, and GPU metrics into sub-sections. Use Tab / Shift+Tab to cycle between sub-sections. Sections without data are dimmed but still accessible. The active sub-section is persisted across sessions.

Key	Action
`Tab` / `Shift+Tab`	Cycle sub-sections (Temperature, Power, ECC, GPU)
`j` / `k`	Navigate sensor/zone list
`Space`	Toggle sensor/zone in chart
`a`	Select / deselect all

Process View

Shows all processes on the system. Non-enriched processes display -- for cmdline and FD count. The history chart shows top N processes by CPU over time.

Key	Action
`j` / `k`	Navigate process list
`c`	Sort by CPU
`m`	Sort by memory
`p`	Sort by PID
`n`	Sort by name
`t`	Sort by threads
`f`	Sort by FDs
`*`	Pin/unpin selected process
`a`	Create alert for selected process
`/`	Search by name or cmdline
`Esc`	Clear search filter
`P`	Toggle pinned-only filter
`Tab`	Toggle history chart: Top CPU / Pinned

Alerts View

Two panels: a rules list on the left and fired alerts on the right. Create, toggle, and delete rules directly from the TUI.

Key	Action
`j` / `k`	Navigate rules or alerts
`Tab`	Switch focus between rules and alerts
`n`	Create new alert rule
`d`	Delete selected rule
`Space`	Toggle rule enabled/disabled
`Enter`	Acknowledge selected alert
`t`	Test all notification channels
`Esc`	Cancel alert creation form

Historical Charts

CPU, memory, disk, hardware (temperature/power/GPU), and process views include a historical braille chart below the live data. Use < / > to cycle through time ranges: 1h, 6h, 24h, 7d, 30d.

Bucket size auto-scales based on the selected range (1 minute for 1h up to 6 hours for 30d). History data is fetched asynchronously and cached per-view for instant display on tab switch.

Staleness Detection

The status bar monitors when fresh data last arrived for the current view. If no new data appears within 3x the longest collector interval for that view, a stale indicator shows: stale (Xs ago).

Debug Mode

bewitch -debug

Adds a scrollable debug console at the bottom of the TUI showing timestamped diagnostic messages: data fetches, cache hits/misses, view transitions, errors, and pin operations.

Key	Action
`{` / `}`	Scroll debug console up / down
`(` / `)`	Shrink / grow debug panel

Layout

The dashboard adapts to a multi-column grid layout on terminals wider than 120 columns.