Searching logs

The Log Explorer is the search-and-filter view for raw log lines from every source. Layout: time picker, filters, paginated table.

Time picker

Quick presets across the top: 15m | 1h | 6h | 24h | 7d. Pick a preset, or set explicit From/To datetimes via the date inputs to the right of the presets.

The time range is the most important filter — narrow it before adding others.

Filters

FilterNotes
Search logs (textbox)Free-text match against the log message body.
All Levels (dropdown)Filter by log level — debug, info, warn, error, trace, unknown.
All namespaces (dropdown)Kubernetes namespaces. Multi-select.
All pods (dropdown)Pod names. Multi-select.
All containers (dropdown)Container names. Multi-select.
Clears all filters at once.

Filters compose with AND. The N results counter above the table updates as you change them.

Results table

Six columns:

ColumnMeans
TimeTimestamp in your local timezone with offset.
LevelPill: DEBUG (gray), INFO (blue), WARN (amber), ERROR (red).
NamespaceKubernetes namespace if applicable.
PodPod name.
ContainerContainer name.
MessageThe log line. Truncated; click the row to see the full message.

Row detail (right-side panel)

Click any row. A panel slides in from the right with:

  • Header — level pill + full timestamp + close (✕).
  • MESSAGE — full message body, including parts truncated in the table. Copy button top-right copies the message to clipboard.
  • DETAILSSource (source ID), Namespace, Pod, Container, Node, Stream (stdout or stderr).
  • LABELS — pills showing every Kubernetes label attached to the producing pod (e.g., app=..., pod-template-hash=...).

The panel stays open as you click other rows — you can quickly compare two lines.

Loading more

The table loads ~50 rows initially. Load more at the bottom pulls the next page. There is no infinite scroll — pagination is intentional so the URL state stays predictable.

Refresh

Refresh top-right re-runs the query. The page does not auto-refresh — refresh manually or use Live Tail for a streaming view.

Exporting

Export CSV top-right downloads the current result set (after filters) as a CSV. The export row cap is set server-side via server.max_export_rows (default 100,000).

Tips

  • For high-volume sources, narrow by namespace or pod before searching the message body. Free-text search across all logs is slower than a level + namespace filter.
  • The 24h preset is enough for most investigations. Use 7d only when you suspect a longer trend.
  • Deep-link from the Overview page — every KPI tile and namespace row links here pre-filtered.