Skip to main content

Compare RRPairs Across Datasets

Speedscale lets you compare request/response pairs across two datasets — snapshots, replay reports, or mock recordings — to identify exactly what changed. This is useful for validating deployments, comparing environments, and catching regressions before they reach production.

Use Cases​

  • Before/after a deploy — capture traffic before and after a code change, then compare to see what responses changed
  • Staging vs. production — compare the same requests across environments to verify staging behaves like production
  • Version comparison — replay the same snapshot against two versions of a service and diff the results
  • Mock validation — compare mock responses against real service responses to verify mock accuracy

Comparing in the Dashboard​

Selecting Two Datasets​

  1. Navigate to the Snapshots or Reports section in the dashboard
  2. Select the first dataset (your baseline)
  3. Select the second dataset (the one you're comparing against)
  4. Click Compare to open the comparison view

Reading the Diff View​

The comparison view shows RRPairs side by side with differences highlighted:

Side-by-side RRPair comparison showing field differences

  • Status code changes — highlighted when the response code differs between datasets
  • Header differences — added, removed, or modified headers are called out
  • Body differences — field-level diffs show exactly which values changed in the response body, with additions in green and removals in red
  • Latency differences — response time changes are shown so you can spot performance regressions

Filtering to Changed RRPairs​

By default, the comparison shows all RRPairs. Use the filter controls to narrow the view:

  • Show only changed — hide matching RRPairs and focus on differences
  • Filter by change type — show only status code changes, body changes, or header changes
  • Filter by endpoint — focus on a specific API path

Comparing via the CLI​

The proxymock CLI includes a files compare command for diffing RRPair files locally when they share refUuid relationships - for example, traffic recorded with proxymock and the replay results generated from it:

proxymock files compare --in recorded/ --in replayed/

This compares RRPair files from the supplied directories or files and exits non-zero when differences are found. Add more verbosity if you want a more detailed diff:

# Compare recorded traffic and replay results with very verbose output
proxymock files compare --in recorded/ --in replayed/ -vvv

Interpreting Diffs​

When reviewing comparison results, focus on these categories:

Status Code Changes​

A status code change (e.g., 200 → 500) is the most obvious regression signal. Investigate these first.

Body Field Changes​

Not all body changes are regressions. Common expected changes include:

  • Timestamps — naturally differ between runs
  • Request IDs and correlation IDs — unique per request
  • Ordering — array elements may appear in a different order

If you expect noisy fields such as timestamps or request IDs, normalize that data before comparing, or use the field ignore controls in the dashboard.

Latency Changes​

A significant latency increase on a specific endpoint may indicate a performance regression, a slower dependency, or infrastructure differences between the two environments.

Header Changes​

New or removed headers can indicate middleware changes, proxy configuration differences, or framework upgrades.

Exporting a Comparison Report​

From the dashboard comparison view, click Export to download the comparison as a report. The export includes:

  • Summary statistics (total RRPairs, changed count, change rate)
  • Per-endpoint breakdown of changes
  • Full diff details for each changed RRPair

This is useful for attaching to pull requests, sharing with QA, or archiving for compliance.