Product / Production comparisons

See what changed after your last release.

Add a release identifier to your traces, compare the sessions that share it through Groups, and inspect the affected evidence before deciding what changed.

compare like with like

24M+ messages ingested

Read the docs ›

Groups

Compare the properties your system actually emits.

The production Groups view can compare sessions by customer, model, environment, or another useful trace property. A release can be one of those properties when you include it in telemetry.

Grouped byrelease.version
4,642 matching sessions · last 30 days
Selected production grouprelease/1.9

1,786 sessions in this production window.

Evidence from release/1.9

Chats behind the difference

Select a version above to change the evidence window

Prepare the Acme briefNo issue detectedSuccessful

Find platform engineersRecovered after one retrySuccessful

Sync qualified leadsThree repeated tool callsNeeds review

Evidence path

What to compare, and where to verify it.

QuestionCompareOpenVerify
Issue-free sessionsCompare by groupIssue detailEvidence sessions
Affected usersCompare by groupUser insightsUser profiles
Execution behaviorCompare by groupSearchFull sessions
MCP outcomesCompare by use caseJourneysTool paths

Honest analysis

A release label is context, not proof.

Traffic mix, environment, and user intent can all change at the same time as a deploy.

VERIFY BEFORE ACTING

Open the sessions behind the difference.

Compare similar users, intents and time windows. Inspect issue details and MCP journeys to investigate the difference; an observational comparison alone does not establish cause.

Evidence first

Use Flowlines with your assistant

Know what changed after a deploy.