Keep a debugging session readable

A busy app can bury the event you are investigating. Filter the event list, keep enough history for the reproduction, and distinguish hidden evidence from evidence the runtime never recorded.

Filters

Click + filter and choose an IN or OUT pattern. IN retains matching events; OUT excludes them. With no IN pills, everything is included before the exclusions are applied.

Pattern	Matches
:standard-epochs/increment-flow	That exact event id
:standard-epochs/*	Every event id in that namespace
:standard-epochs	That id and ids in that namespace
increment	Any event id containing that text

Edit a pattern with its pencil control or remove it with ✕. Right-click an event and choose Always hide this event-type… to add an exclusion, or Mute to hide the id. The ribbon's 🔇 N opens the mute manager. Pattern pills reset on reload unless the host supplies a boot-time filter seed. Remove both pills and mutes when checking whether an event happened.

An errored event remains visible even when a filter or mute would hide it, with a tooltip explaining the override. Warnings do not override filters. This behaviour is configurable in the filter reference.

Managed-effect context menus also create filters for an effect id or one HTTP correlation id. These include the matching activity's spawning ancestors so you can still find the user action that caused it. They do not resurrect evicted evidence.

History and settings

Epoch history (3) retains successful and failed event snapshots per frame. The separate Buffer tab sets the number of events kept in the framework's trace ring. Both default to 50. Increase the relevant setting before reproducing a long interaction; existing gaps stay gaps.

Clear buffer now asks for confirmation and drops retained evidence and the redaction counter. Clear at the beginning of a new investigation if needed, then reproduce the problem. Hiding Xray or pausing following does not clear the history.

Settings are saved in the browser. A developer's explicit choice takes precedence over the project's configured defaults on the next load. The configuration reference documents the precedence and exact values.

Redacted and large values

By default Xray uses :rf.egress/local-redacted: classified sensitive values display as markers, and sensitive trace events may be held back. ● REDACTED N reports the number held back. This is separate from an IN or OUT filter.

For trusted local inspection, configure the local raw profile:

(require '[day8.re-frame2-xray.core :as xray])
(xray/configure! {:rf.xray/egress-profile :rf.egress/local-raw})

This changes Xray's local visibility, including sensitive and large values; it does not change Story's or another tool's profile. Returning to :rf.egress/local-redacted clears the trace buffer, so previously visible raw payloads do not remain there. Choose the profile before reproducing the interaction. A value's source link can remain useful while its value is hidden.

Troubleshooting

Symptom	Check	Action
The list is empty	Frame, IN/OUT pills and mutes	Choose a live app frame and remove exclusions
New events are recorded but the panel stays still	Focus or paused following	Press » or resume with Space
An error remains despite an OUT filter	Error override is enabled	Inspect it, or change that setting for this investigation
An outside-event error is absent	Ungrouped rows are hidden by default	Enable the ungrouped option in General settings
A setting changes back after reload	Persisted override or host defaults	Check the documented precedence rather than changing unrelated code
History disappeared after changing privacy	The raw-to-redacted transition clears it	Reproduce under the new profile