Replace O(windows × sites) isMonitorInMaintenance with O(1) map lookup.
buildMaintSet runs once per refreshLive/handleTabData, not per call site.
groupSparkline/groupUptime now use m.sites (already on model) instead of
calling engine.GetAllSites() which copies the full map under a mutex.
Replace the cramped 22-char maint sidebar with a full-width bottom
strip in the same position as logs. Bottom panel is modal: l for logs,
m for maint, same key again to close. Maint strip shows one row per
window with icon, title, affected monitors, and status/time remaining.
Removes maintSidebarW constant and maintOffset field. Introduces
bottomPanel enum (bottomNone/bottomLogs/bottomMaint) replacing
logsOpen and maintOpen booleans. m key no longer gated behind
wideBreakpoint.
Detail sidebar now maintains consistent height matching the monitors
panel across all modes (default, SLA, history). Content scrolls within
the fixed frame via mouse wheel or j/k when the detail panel is
focused. Scroll offset resets on mode/monitor/period changes. Logs
strip stays visible in all states.
All monitor detail content now lives in the sidebar panel.
Delete the full-page viewDetailPanel, viewSLAPanel, viewHistoryPanel,
handleDetailKey, handleSLAKey, handleHistoryKey, openSLAView, and
related dead code (histogram, sparkline tooltip, detailViewport).
Remove stateDetail, stateSLA, stateHistory session states.
Update tests to use detailOpen/detailMode instead of removed states.
SLA and History views now render inside the detail sidebar panel
instead of taking over the full screen. Press s/h to switch modes,
Esc to return to default detail view. Cursor movement resets to
default mode.
- Add detailMode state (detailDefault/detailSLA/detailHistory)
- SLA sidebar: period selector, uptime bars, daily breakdown
- History sidebar: sparkline density, transition list, stats
- Panel title shows "SLA · name" or "History · name" in sub-modes
- Period keys [1]-[4] work in SLA mode from dashboard state
Replace 3-tab navigation with a single monitoring dashboard.
- Replace tab bar with minimal status line (pulse + counts + version)
- Maintenance becomes a toggleable left sidebar (m key)
- Compact list: active (●) and scheduled (○) windows
- Enter opens full-screen detail with end/delete actions
- Settings becomes a centered floating overlay (S key)
- Alerts/Nodes/Users sub-sections with left/right navigation
- Full CRUD support inside overlay with returnState tracking
- Remove currentTab, switchTab, tab constants, tab click zones
- Replace deleteTab with deleteKind string for cleaner dispatch
- Separate settingsCursor/settingsOffset from main cursor
- Panel focus: maint → monitors → logs → detail
- Context-sensitive footer key hints per focused panel
- Click Alerts/Nodes/Users sub-tabs in settings to switch sections
- Click STATUS/NAME/LATENCY column headers to sort (click again
to reverse direction)
- Logs panel show/hide: [l] toggles sidebar visibility, click
focused panel to hide. Monitors expand to full width when hidden.
- Logs full-screen: Enter when focused on logs opens scrollable
full-screen log view with severity tags, filter toggle [f],
and live updates
Thread context.Context from callers instead of creating
context.Background() at every call site. Engine stores its
lifecycle ctx for use by triggerAlert goroutines. TUI Model
carries ctx for store operations in Cmd closures. CLI commands
create a root ctx and thread it through openStore, seed
functions, and init methods.
Only two intentional context.Background() remain in non-root
positions: engine constructor default (overridden by Start)
and drainWrites (parent already cancelled at shutdown).
Refactored full-screen detail panel: status+endpoint on left column,
timing+config on right column. Halves vertical height of the info
section. Content wraps in a viewport when it exceeds terminal height
— scroll with j/k/arrows/pgup/pgdn.
Restored renderSparkTooltip lost in rewrite. Fixed lint warnings.
Statuspage-style colored bar — one █ per day, colored by uptime:
green (>= 99%), yellow (>= 95%), red (< 95%), gray (no data).
30-day percentage shown to the right.
Daily breakdown computed via ComputeDailyBreakdown from state_changes
and loaded alongside detail data in loadDetailCmd. Auto-updates on
cursor move.
Press < / > to cycle sort column (Status, Name, Latency). Press r to
reverse direction. Sorted column shows ▲/▼ arrow in the header.
Groups always float to top. Sort applies to ungrouped monitors and
group children independently. Default: Status descending (DOWN first).
Detail open/closed state saved via SetPreference on toggle and
restored on session start. Same pattern as theme persistence —
survives restarts and works per-user over SSH.
Click any panel (Monitors, Logs, Detail) to focus it — accent border
follows focus. Mouse wheel scrolls the focused panel.
Keyboard: l toggles log panel focus. Arrow keys scroll logs when
focused, navigate monitors when not. Esc returns focus to monitors.
Log sidebar now supports scroll offset — tracks position across
renders without a viewport. Mouse wheel scrolls 3 lines, keyboard
scrolls 1.
Press i to toggle a compact detail panel below the monitors+logs
split. Shows status, latency, uptime, state changes, sparkline, and
key hints in ~6 lines. Auto-updates when cursor moves between
monitors. h/s/e keys work from the inline detail for history, SLA,
and edit. Escape closes the panel.
No more full-screen detail takeover for the common case. The old
stateDetail path remains for h/s sub-views which still go full-screen.
Table columns were computed from terminal width, causing row wrapping
when the monitors panel only gets 70% of the space. Introduced
contentWidth field set per-tab in viewDashboard. computeLayout,
isWide, and renderTable now use contentWidth for column visibility,
available space, and max table width calculations.
Columns gracefully hide (SSL, RETRIES, TYPE, UPTIME) when the panel
is narrower, matching the existing responsive breakpoint behavior.
Tab bar: Monitors | Maint | Settings (was 6 tabs).
Settings tab merges Alerts, Nodes, Users as sub-sections with
left/right arrow navigation. Each section keeps its own cursor,
keybindings, and CRUD operations.
Monitors tab now shows a log sidebar at >= 120 cols (70/30 split).
Under 120 cols, monitors render full-width without logs.
- Introduced tab constants (tabMonitors, tabMaint, tabSettings)
- Introduced section constants (sectionAlerts, sectionNodes, sectionUsers)
- Removed stateLogs and stateUsers states
- All magic tab numbers replaced with named constants
Apply Bold/Faint attributes to semantic styles following htop's
monochrome design principle. Creates 4-tier visual hierarchy that
works even when colors collapse: Bold (danger/warn), Normal (success/
default), Faint (subtle/stale/borders/inactive tabs). Complements
the ANSI-16 color fallbacks without affecting TrueColor appearance.
Theme colors now use lipgloss.CompleteColor with hand-picked ANSI-16
values instead of raw hex. Prevents algorithmic degradation from
collapsing dark backgrounds into indistinguishable ANSI colors over
SSH. Backgrounds fall through to terminal default in 16-color mode;
semantic colors map to distinct ANSI indices (green/yellow/red/blue/
cyan/magenta). TrueColor rendering is unchanged.
Replace 4-page paginated form (17 fields for HTTP) with a 2-page
type-aware layout. Page 1 shows core fields + type-specific target
(URL for HTTP, Hostname for ping, etc). Page 2 shows configuration
with pre-filled defaults. Group type gets 1 page.
Form rebuilds dynamically when monitor type changes, preserving
all entered values via pointer-bound siteFormData. Focus returns
to the Type select after rebuild so users can continue forward.
WithWidth set explicitly on rebuild to prevent placeholder truncation.
viewLogsTab filtered logViewport.View() — the visible window — so the
entry count showed the window size and hidden lines reappeared while
scrolling. Filter and render now happen at content-set time from
engine.GetLogs(); the view only reads stored counts.
1. Delete braille.go + braille_test.go — dead code, only referenced
by its own test. Can be re-added when latency charts are built.
2. Hoist duplicate `const sparkWidth = 40` (update.go + view_detail.go)
to package-level `detailSparkWidth`. Click-index resolution and
rendering now share one constant.
3. Remove tea.ClearScreen on every resize — caused full-screen flash
during continuous resizes. ctrl+l manual clear kept.
Cursor tracked by site ID instead of positional index. When the
list re-sorts every tick (sites change status), the selection stays
on the same monitor instead of silently jumping to whatever now
occupies that index position.
q now means "back" in detail, history, SLA, and alert-detail views
— consistent with muscle memory from navigating deeper views.
Only the dashboard q quits the app. ctrl+c always quits from
anywhere.
Every Store interface method (except Close) now takes context.Context
as first parameter. All 54 db.Query/Exec/QueryRow calls in SQLStore
replaced with their *Context variants. DB operations now respect
cancellation and deadlines.
Context sources by caller:
- Engine dbWriter/poll/pruner: engine ctx from Start()
- HTTP handlers: r.Context()
- config.Apply/Export: caller-provided ctx
- TUI/main.go init: context.Background()
RunCheck and all sub-checks (HTTP/ping/port/DNS) accept parent ctx.
HTTP checks now inherit shutdown cancellation instead of rooting in
context.Background(). dbWrite.exec takes ctx so the writer goroutine
can cancel stuck DB operations.
DeleteSite/ImportData use BeginTx(ctx) instead of Begin().
The #101 refactor stopped at the tick path; 'h' history and the SLA
view still queried state changes synchronously in Update, freezing the
UI for up to busy_timeout on a contended DB. Both now load through
Cmds with loading placeholders.
Also closes the remaining staleness holes in the async data flow:
- tabDataMsg carries a sequence number; out-of-order replies from
slower earlier loads are dropped instead of overwriting newer data
- history/SLA replies are dropped when the user has navigated to a
different site or period
- the open detail panel refreshes on the tab-data cadence instead of
loading once on entry and going stale
- initSiteHuhForm reads the m.alerts cache instead of hitting the store
applyTheme mutated ~18 package-global lipgloss styles while every SSH
session's tea.Program read them concurrently from its own goroutine.
Pressing T or opening a new connection raced other sessions' View and
bled themes across users.
Styles now live in an immutable per-Model struct built by newStyles;
free formatter helpers that consumed the globals became Model methods.
The TUI ran database queries on the UI goroutine: handleTick called
refreshData every second, which issued four blocking SQLite queries
(GetAllAlerts/GetAllUsers/GetAllNodes/GetAllMaintenanceWindows) and
swallowed their errors; viewDetailPanel ran GetStateChanges — a DB query
— inside View(), on every render (tick, keypress, mouse). A slow disk
stalled input and animation.
Split refreshData into refreshLive() (in-memory engine copies only —
sites + logs — safe every tick) and loadTabDataCmd(), a tea.Cmd that
loads the four DB-backed tables off the UI goroutine and returns a
tabDataMsg. handleTick now refreshes live state every tick but dispatches
the tab-data load only when older than tabRefreshTTL (5s), so tab-bar
counts stay fresh without a per-second query storm. Errors surface to the
log instead of being dropped, and a transient failure keeps the previous
data rather than blanking the view.
The detail panel's state-change history is loaded once on enter via
loadDetailCmd and cached on the model; viewDetailPanel reads the cache,
so View no longer touches the database. Init kicks an initial load so the
dashboard isn't empty on the first frame, and the bare time.Time tick
message is now a named tickMsg (no cross-message collision). The
test-alert handler's raw goroutine becomes a tea.Cmd.
Adds the package's first Update()-driven tests: tab-data load + apply,
error-keeps-previous-data, detail cache with a store-hit counter proving
View does zero IO across repeated renders, and the handleTick throttle.
Full suite green under -race; golangci-lint clean.
Click any sparkline character to see data point details — approximate
time, latency, and up/down status. Esc dismisses tooltip without
leaving detail view. Uses existing BubbleZone infrastructure with
zone-relative coordinate math for index resolution.
Replace misleading relative-only sparkline with dual-channel design:
bar height uses relative scaling (shows stability and anomalies),
color+brightness uses absolute thresholds (shows fast vs slow).
- Add brightness gradient within color bands (dim→bright as latency
increases toward the next threshold)
- Pass row background through sparkline rendering so zebra stripes
and selection highlights carry through ANSI sequences
- Cap sparkline width to 60 (matches maxHistoryLen) and column
width to 62 to eliminate trailing dead space
- Quiet group sparkline: subtle dots for healthy, bold red for down
- Add braille subpixel canvas (ported from meridian) for future
multi-row graph use
- Show version in dashboard footer (wired from goreleaser ldflags)
- Cap name column at 35, raise sparkline minimum to 15 chars
- Preserve zebra background on group rows (was lost by style override)
- Group sparkline uses bullet • instead of heavy circle ●
Full-screen SLA report accessible via [s] from detail panel.
Computes uptime%, downtime, outage count, longest outage, MTTR,
and MTBF from state_changes table. Includes daily breakdown with
bar chart, switchable time periods (24h/7d/30d/90d), and
scrollable viewport. LATE/STALE treated as UP for SLA purposes.
New intermediate state between LATE and DOWN at the midpoint of
the grace period. Gives operators earlier warning that a push
monitor has gone quiet. Includes dedicated orange theme color
across all 5 themes and proper styling in dashboard, detail
panel, and history view.
Full-screen scrollable history view accessible via [h] from detail
panel. Shows all state transitions with computed outage durations,
event density sparkline for flapping detection, and summary stats.
- Detail panel STATE CHANGES now shows outage duration per recovery
- Event density sparkline highlights flapping periods
- Summary footer: event count, outage count, avg outage duration
- Vim-style navigation (j/k/g/G) + mouse scroll in history view
tui.go (1032→164) and tab_sites.go (993→482) violated "small functions"
and "testable in isolation" standards. Extracted 6 new files by concern:
- format.go: pure formatting functions (fmtLatency, fmtUptime, etc.)
- sparkline.go: sparkline rendering (latency, heartbeat, group)
- update.go: Update method decomposed into 15 named handlers
- view_dashboard.go: View, dashboard composition, tab bar, footer
- view_detail.go: site detail panel
- data.go: data refresh with extracted sortSitesForDisplay/filterSites
Added 17 unit tests for the newly-testable pure functions covering
format, sparkline, sort ordering, and filter logic. No behavioral
changes — strict move-and-extract refactor.
Move Go module from gitea.lerkolabs.com/lerko/uptop to
gitea.lerkolabs.com/lerkolabs/uptop. Updates all imports,
go.mod, goreleaser owner, and README links.
- Add 6 TUI screenshots to assets/ (monitors, alerts, logs, nodes, detail, theme)
- Rewrite README with hero image, badges, collapsible install sections
- Rewrite changelog to match actual CalVer tag history
- VHS tooling extracted to lerko/uptop-vhs
Reviewed-on: lerko/uptop#32
Track alert delivery health at runtime:
- AlertHealth struct: LastSendAt, LastSendOK, LastError, SendCount, FailCount
- triggerAlert records success/failure after each Send()
- Health data exposed via GetAlertHealth() for TUI
Alerts tab enriched:
- Health dot column: green (OK), red (failed), gray (never sent)
- LAST SENT column: relative time ("2m ago", "never")
- [t] key sends test notification through selected channel
Inspired by Grafana's contact point health columns.
Push monitors no longer lie about status:
- PENDING stays until first heartbeat (no auto-promote to UP)
- LATE state (amber) when overdue but within grace period
- DOWN only after grace period expires
- Grace period = interval/2, minimum 60s
RecordHeartbeat now handles all transitions:
- PENDING → UP (first heartbeat, logged)
- LATE → UP (late arrival, logged)
- DOWN → UP (recovery, alert + state change persisted)
TUI updates:
- LATE rendered in amber/warning color
- Status bar shows LATE count separately
- Tab badge shows ⚠ for late monitors
- Sort order: DOWN > LATE > UP > PENDING > PAUSED
- Detail panel shows error for LATE monitors
Inspired by Healthchecks.io state machine (new/up/grace/down).
Flexoki Dark (default), Flexoki Light, Catppuccin Mocha, Nord.
Press T to cycle themes; selection persists in preferences.
All hardcoded colors replaced with theme-driven values.
Dedicated ZebraBg per theme for subtle row striping.
Site count in tab label and footer now reflects total monitors
(excluding groups) regardless of collapse state. Down count also
computed from all sites so collapsed groups with down children
still surface in the badge. Replaced Nerd Font folder glyphs
with standard Unicode triangles for cross-font compatibility.
Add alternating row backgrounds for easier table scanning. Detail panel
now shows breadcrumb path (Sites > Group > Name) and min/avg/max latency
stats below the sparkline. Group collapse state persists across restarts
via new preferences table in both SQLite and Postgres.
Group rows now show selection background when navigated to. Layout
chrome extracted to named constants to prevent viewport drift. Groups
display aggregate history as dot sparkline (●) distinct from site
bar sparklines, with uptime computed from active children only.
Paused and maintenance children excluded from group aggregates.
Forms overflowed past terminal because huh didn't know about the
surrounding chrome (header, footer, padding). Now sets WithHeight()
on every render and forwards WindowSizeMsg during form state.
lerko
