Run Logs — SideButton Dashboard Knowledge Module

What This Is

The Run Logs module covers three related views for monitoring and inspecting workflow executions: the Run Log List (history + live running workflows with KPI stats), the Run Log Detail (full event timeline for a completed run), and the Execution View (real-time streaming log for an in-progress execution). Together they form the execution observability surface of the SideButton dashboard.

URL Patterns

View	URL	Notes
Run Log List	/run-logs	No sidebar nav entry — reached via Skills "View Run Logs" link or other in-app navigation
Run Log Detail	/run-logs/{id}	id is the run log record ID (not run_id)
Execution View	No URL	Internal view state only — triggered by navigateToExecution(workflowId, runId). Cannot be deep-linked

---

View 1: Run Log List (RunLogView.svelte)

Page Structure

+--[Header: "Run Logs" h1 + item count badge]--[Clear All] [Reload]--+
|  [.kpi-bar] (only when runs exist)                                  |
|    [.kpi-hero: time saved today value + label]                      |
|    [.kpi-time-breakdown: This Week / This Month / All Time rows]    |
|    [.kpi-divider]                                                   |
|    [.kpi-stats: Runs Today + Success Rate]                          |
|                                                                      |
|  [Currently Running section] (only when running workflows exist)    |
|    h2 with .pulse dot + "Currently Running ({count})"               |
|    [.running-list: .running-card buttons]                           |
|                                                                      |
|  [.history-section]                                                  |
|    h2 "History" (only shown when running section is also visible)   |
|    [.run-list: .run-card buttons]                                   |
|                                                                      |
|  [.empty-state] (when no runs and not loading)                      |
|  [.loading-state] (while fetching)                                  |
+----------------------------------------------------------------------+

Key Elements

Header

Element	Selector	Notes
Page heading	h1 containing "Run Logs"	Inside .header-left
Item count badge	.item-count	Shows runs.length
Clear All button	.btn-secondary in header	Initiates inline confirm flow — does NOT open a dialog
Reload button	.reload-btn	Refetches run logs and running workflows immediately

Clear All Inline Confirm

Clicking "Clear All" does NOT use window.confirm(). It replaces the button area with an inline confirm UI in-place.

Element	Selector	Notes
Confirm text	.confirm-text	Text: "Clear all logs?"
Confirm delete button	.btn-danger	Text: "Yes, Clear". Executes the clear operation
Cancel button	.btn-secondary (inside confirm)	Text: "Cancel". Dismisses confirm without action

KPI Stats Bar

Visible only when runs.length > 0. Hidden entirely during empty state.

Element	Selector	Notes
Stats bar container	.kpi-bar	Horizontal row; absent when no runs
Hero value	.kpi-hero-value	Large green text — time saved today (formatted duration string)
Hero label	.kpi-hero-label	Text: "Time Saved Today"
Time breakdown rows	.kpi-time-breakdown	Contains rows for This Week, This Month, All Time
Time row label	.kpi-time-label	e.g., "This Week", "This Month", "All Time"
Time row value	.kpi-time-value	Formatted duration for that period
Vertical divider	.kpi-divider	Separates hero + breakdown from stats column
Stats column	.kpi-stats	Contains Runs Today count + Success Rate percentage

Running Workflows Section

Shown only when $runningWorkflows.length > 0.

Element	Selector	Notes
Section heading	h2 with .pulse dot	Animated dot + "Currently Running ({count})" text
Running list	.running-list	Flex/grid container of running cards
Running card	button.running-card	Clickable — calls navigateToRunningExecution(run_id, workflow_id) → opens ExecutionView
Card header	.running-header	Contains workflow title + StatusBadge showing "running"
Card meta	.running-meta	Shows run_id + started time

History Section

Element	Selector	Notes
History section container	.history-section	Always rendered when runs exist
History heading	h2 "History"	Only rendered when the Running section is also visible (i.e., hasRunning is true)
Run list	.run-list	Container of completed run cards
Run card	button.run-card	Clickable — calls navigateToRunLogDetail(run.id) → opens Run Log Detail
Card header	.run-header	Contains workflow title + StatusBadge
Card meta	.run-meta	Shows run_id + duration + timestamp

StatusBadge Component

Used on both running cards and history cards.

Status value	Visual
"running"	Blue badge (animated)
"success"	Green badge
"failed"	Red badge

Empty and Loading States

Element	Selector	Notes
Empty state	.empty-state	Clock SVG icon + h3 "No Run History" + "Workflow executions will appear here."
Loading state	.loading-state	Text: "Loading run logs..."

WebSocket Behavior

• Subscribes to running-workflows-changed events on mount
• When a running workflow completes, triggers auto-refresh of run logs after a 500ms delay
• Unsubscribes on component destroy

Data Sources

• listRunLogs(50) — fetches up to 50 most recent completed run logs
• getRunningWorkflows() — fetches currently active executions

---

View 2: Run Log Detail (RunLogDetailView.svelte)

Page Structure

+--[Back btn] [.header-actions: Delete confirm area]--+
|  [.run-info container, max-width 900px]              |
|    [.title-row: h1 workflow title + StatusBadge]     |
|    [.meta-grid: auto-fill grid of .meta-item cards]  |
|    [.params-list] (conditional: when params exist)   |
|    [h2 "Events ({count})"]                           |
|    [.events-container: .event-item rows]             |
|                                                      |
|  [.not-found] (when ID not found)                   |
|  [.error] (on API error)                             |
|  [.loading] (while fetching)                         |
+------------------------------------------------------+

Key Elements

Header

Element	Selector	Notes
Back button	.back-btn	Text: "Back". Calls navigateToRunLogs()
Header actions area	.header-actions	Right side of header — contains delete button or inline confirm
Delete button	button inside .header-actions	Initial state — triggers inline confirm

Delete Inline Confirm

Clicking Delete replaces the button with inline confirm UI in-place. Does NOT use window.confirm().

Element	Selector	Notes
Confirm text	.confirm-text	Text: "Delete this log?"
Confirm delete button	.btn-danger	Text: "Yes, Delete". Deletes and navigates to run logs list
Cancel button	.btn-secondary	Text: "Cancel". Returns to delete button

After delete completes: navigateToRunLogs() is called automatically.

Run Info

Element	Selector	Notes
Info container	.run-info	max-width: 900px, centered
Title row	.title-row	Contains h1 (workflow title) + StatusBadge component
Meta grid	.meta-grid	grid-template-columns: repeat(auto-fill, minmax(..., 1fr))
Meta item card	.meta-item	One card per metadata field
Meta label	.meta-item .label	Uppercase text — field name (e.g., "RUN ID", "DURATION")
Meta value	.meta-item .value	Field value
Monospace value	.meta-item .value.mono	Used for IDs and timestamps

Meta grid fields (one card each): Run ID, Workflow ID, Duration, Events, Triggered By, Timestamp.

Parameters Section

Shown only when runLog.params has one or more entries.

Element	Selector	Notes
Params list	.params-list	Grid of parameter cards
Param card	.param-item	One per parameter
Param name	.param-name	Purple text — parameter key
Param value	.param-value	Monospace — parameter value

Events Section

Element	Selector	Notes
Section heading	h2	Text: "Events ({count})"
Events container	.events-container	Monospace font family
Event row	.event-item	Depth-based left padding (calc(depth * N px))
Event index	.event-index	Row number (0-based)
Event type	.event-type	Type label with color class (see below)
Event message	.event-message	Content of the event

Event Type CSS Classes

Event type	CSS class	Color
workflow_start	.event-run	Blue
workflow_end	.event-end	Green
step_start	.event-step	Gray
step_end	.event-step-end	Purple
error	.event-error	Red
log	.event-log	Orange / warning

Terminal States

Element	Selector	Notes
Not found	.not-found	Text: "Run log not found."
Error	.error	API error message text
Loading	.loading	Text: "Loading run log..."

---

View 3: Execution View (ExecutionView.svelte)

Page Structure

+--[.back-btn (disabled while running)] [.header-info] [.header-actions]--+
|  [.error-banner] (on error)                                               |
|                                                                            |
|  [.log-header: event count + run ID]                                       |
|  [.log-container: scrolling monospace log area]                            |
|    [.empty-state with .spinner] (while waiting for first event)            |
|    -- or --                                                                |
|    [.log-entry rows]                                                       |
+----------------------------------------------------------------------------+

Key Elements

Header

Element	Selector	Notes
Back button	.back-btn	Disabled (disabled attr) while $isRunning. Enabled when completed
Header info	.header-info	Contains h1 (workflow title or "Execution" if title unknown) + status badge
Status badge (running)	.status.running	Blue — contains .pulse animated dot
Status badge (completed)	.status.completed	Green
Header actions	.header-actions	Right side — contains Stop or Done button
Stop button	.btn-stop	Red-tinted. Visible while $isRunning. Calls stopWorkflow(runId)
Done button	.btn-primary	Visible when completed. Navigates back

Error Banner

Element	Selector	Notes
Error banner	.error-banner	Shown when an error occurs. Contains error message text

Log Area

Element	Selector	Notes
Log header	.log-header	Shows event count + run ID
Log container	.log-container	Monospace font, scrollable. Auto-scrolls to bottom as new entries arrive
Empty state	.empty-state	Shown when $logs.length === 0. Contains .spinner + "Waiting for execution events..."
Spinner	.spinner	Spinning blue circle (CSS animation)
Log entry	.log-entry	One row per event. Depth-based left padding
Log time	.log-time	Timestamp column
Log type	.log-type	Type label with color class (see below)
Log message	.log-message	Content text

Log Entry Type CSS Classes

Event type	CSS class	Color
workflow_start	.log-run	Blue
workflow_end (success)	.log-success	Green
error	.log-error	Red
step_start / step_end	.log-step	Gray
result / step_end with result	.log-result	Purple
log with level warn	.log-warn	Orange

WebSocket Behavior

• Calls initWebSocket() on mount — opens real-time connection for log event streaming
• $effect watches $logs.length — auto-scrolls .log-container to bottom on every new entry
• stopWorkflow(runId) stops the execution and sets $isRunning to false
• Back button remains disabled (disabled attribute) while $isRunning === true

---

Data Models

RunLogMetadata

Field	Type	Notes
id	string	Run log record ID (used in /run-logs/{id} URL)
workflow_id	string	Workflow identifier
workflow_title	string	Display name
status	string	"success", "failed", or "running"
duration_ms	number	Total execution time in milliseconds
event_count	number	Number of events recorded
timestamp	string	ISO 8601 datetime of execution start
triggered_by	string	Origin: "user", "agent", etc.

RunLog (full, fetched for detail view)

Field	Type	Notes
metadata	RunLogMetadata	See above
events	WorkflowEvent[]	Complete ordered event list
params	Record<string, string>	Input parameters provided at execution time

WorkflowEvent

Field	Type	Notes
type	string	workflow_start, workflow_end, step_start, step_end, error, log
action_id	string	Workflow ID this event belongs to
action_title	string	Workflow display name
step_index	number	0-based step position
step_type	string	Step type name
step_details	string	Step detail text
success	boolean	Outcome — present on workflow_end and step_end events
message	string	Error or log message text
result	string	Step result text (shown as .log-result / .event-step-end)
level	string	For log events: "info", "warn", "error"
depth	number	Indentation level — drives left padding in both detail and execution views
timestamp	string	ISO 8601 datetime of this specific event

RunningWorkflow

Field	Type	Notes
run_id	string	Active run ID (used to navigate to ExecutionView)
workflow_id	string	Workflow identifier
workflow_title	string	Display name
started_at	string	ISO 8601 datetime when execution started

---

States

RunLogView States

State	Trigger	Visual
Loading	Initial fetch in progress	.loading-state "Loading run logs..."
Empty	Fetched, no runs found	.empty-state with clock icon, h3 "No Run History"
Populated (no running)	Runs exist, none active	.kpi-bar + .history-section only
Populated (with running)	Runs exist + active runs	.kpi-bar + Running section + h2 "History" + history section
Clear All: confirm pending	Clicked "Clear All"	.confirm-text + .btn-danger + .btn-secondary replace button
Auto-refresh	Running workflow completes	500ms delay then data refetched silently

RunLogDetailView States

State	Trigger	Visual
Loading	Fetching run log	.loading "Loading run log..."
Loaded	Fetch succeeds	.run-info with title, meta grid, optional params, events
Not Found	ID not in records	.not-found "Run log not found."
Error	API failure	.error with error message
Delete: confirm pending	Clicked Delete	.confirm-text + .btn-danger + .btn-secondary
Deleted	Confirmed delete	Navigates to run logs list

ExecutionView States

State	Trigger	Visual
Running (no events yet)	Started, WebSocket connected, no events	.empty-state with .spinner + "Waiting for execution events..."
Running (events streaming)	Events arriving via WebSocket	.log-entry rows; auto-scrolling; Back disabled; Stop button visible
Completed	Final event received	.status.completed badge; Done button; Back enabled
Error	Error event or API failure	.error-banner visible
Stopped	User clicked Stop	stopWorkflow() called; $isRunning set false; transitions to completed state

---

Common Tasks

1. View run history: Navigate to /run-logs — history cards appear in .run-list sorted most-recent-first
2. Open a completed run: Click a .run-card in the history section → navigates to /run-logs/{id}
3. Open a live execution: Click a .running-card in the Running section → opens ExecutionView with real-time log stream
4. Read KPI stats: Load /run-logs with existing runs — .kpi-hero-value shows time saved today; .kpi-stats shows Runs Today count and Success Rate
5. Delete a run log: In Run Log Detail, click Delete button → .confirm-text appears → click "Yes, Delete" → returns to list
6. Clear all logs: In Run Log List, click "Clear All" → confirm text appears → click "Yes, Clear" → all logs deleted
7. Reload logs: Click .reload-btn — immediately refetches both run logs and running workflows
8. Stop a running execution: In ExecutionView, click .btn-stop → stopWorkflow(runId) called → transitions to completed state
9. Read event details: In Run Log Detail, inspect .event-item rows — color-coded by .event-* class; indented by depth
10. Read parameters: In Run Log Detail, .params-list shows input params if any were provided (absent if no params)

---

Tips

• Running vs History are separate data sources: Running workflows come from getRunningWorkflows() (live store); history comes from listRunLogs(50). They render in separate sections and refresh independently
• History heading is conditional: The h2 "History" heading only renders when there are also running workflows visible. When there are no running workflows, the list renders without a heading
• KPI bar is data-dependent: .kpi-bar is absent from the DOM when runs.length === 0. Do not assert its presence in empty-state tests
• Depth field drives indentation: Both .event-item (detail) and .log-entry (execution) use depth to calculate left padding — deeper nesting = more indentation
• ExecutionView has no URL: It is rendered as an internal view state — navigating directly to any URL will not open it. It must be reached via a running-card click or navigateToExecution() call
• Auto-scroll is reactive: ExecutionView scrolls .log-container to the bottom on every new log entry via a Svelte $effect. The container is always showing the latest event during a live run
• id vs run_id: RunLogMetadata.id is the database record ID used for the detail URL. run_id (on RunningWorkflow) is the active execution handle. These are distinct fields with different uses
• WebSocket refresh delay: When a running workflow completes, RunLogView waits 500ms before refetching — the history list may briefly not show the just-completed run

---

Gotchas

• Clear All and Delete use INLINE confirm, not window.confirm(): Clicking these buttons does NOT open a browser dialog. They transform the button area into .confirm-text + .btn-danger + .btn-secondary in-place. Automation must click through the two-step sequence
• ExecutionView Back button is disabled while running: .back-btn has the disabled attribute while $isRunning. Attempting to click it during execution will have no effect — wait for completion or click Stop first
• No deep link to ExecutionView: navigateToExecution() sets internal view state only. There is no URL to construct. Cannot be reached by navigating the browser address bar
• Run Log List has no sidebar nav entry: /run-logs is not listed in the left sidebar. It must be reached from Skills "View Run Logs" or programmatic navigation — do not expect a sidebar button
• listRunLogs(50) cap: Only the 50 most recent logs are fetched. Older entries are not paginated — they are simply omitted from the list view
• Empty event type falls back: If an event's type does not match a known class (.event-run, .event-end, etc.), no color class is applied — the row renders in default text color
• Params section absent when no params: .params-list is not rendered at all when runLog.params is empty or undefined — do not assert its presence unless params are known to exist
• WebSocket in ExecutionView is separate from RunLogView WebSocket: Each view manages its own WebSocket subscription. The execution view streams live events; the list view only subscribes to running-workflows-changed for refresh triggers
• SPA 404 on direct /run-logs navigation: As with all routes, loading /run-logs directly in the browser address bar before the SPA shell is loaded returns a Fastify 404 JSON response. Must navigate through the SPA root first