A performance-test screenshot should let a reviewer understand the measurement while keeping unrelated information out of the attachment. Work on a copy, retain the labels needed to interpret the graph, and remove confidential details before exporting the image. Keep the original test evidence in its approved location; the screenshot is a view of that evidence, not a replacement for the results.
Start with the question under review. A chart used to discuss a login slowdown needs a different frame from one showing throughput during a load ramp. Decide which observation the image supports before adding arrows or cutting away panels.
Keep the measurement interpretable
Include the metric name, units and time window. If the discussion concerns a percentile, leave its label visible instead of presenting it as an average. Retain the legend when several services or test runs share the plot. A dramatic curve with its scale cropped off gives the reviewer too little information to judge the change.
Put the run identifier and relevant test conditions in the accompanying note. For a comparison, explain whether the workload and environment were held constant; record any differences that could affect the result. If the full report is available to the reviewer, link to its approved location. Do not make a screenshot carry the entire test specification.
Remove private details from a working copy
Inspect the frame beyond the chart. Dashboard filters, browser tabs and expanded request details may expose account names or internal addresses. Crop an unrelated panel when it contributes nothing to the review. When the surrounding layout matters, the practical workflow for redacting personal information from a screenshot explains opaque covers, flattened output and checks on the delivered file.
Use a solid, fully opaque fill over information that must be withheld, extending the cover past every character. Avoid treating a softened blur as equivalent to removing confidential text. Flameshot’s privacy guidance for screenshot tools warns about blur and small-scale pixelation and identifies a filled rectangle as the safer option. Review the whole frame again for repeated values, including those in small labels.
Keep the editable working file separate from the attachment. TechSmith distinguishes SNAG/SNAGX source graphics from flattened PNG and JPEG images: the source format retains editable objects, while those exported image formats incorporate the objects into the image. Renaming a project file does not perform that export. Flattening also does not establish that every metadata field or unrelated private detail has been removed.
Review the attachment alongside its note
Open the exported image outside the editor and examine it at its actual pixel size. Check the edges of each cover and confirm that the metric labels remain readable. Inspect the file’s metadata using an approved local tool when the disclosure requires it, and check that the sharing method does not include the unredacted original or editable project.
Keep annotations away from the axes and redaction masks. A callout can identify the changed curve or a time interval; it should not assert a root cause that the graph cannot establish. The SnagitPro screenshot guides cover capture and annotation choices if the report needs a separate close-up. Give that close-up its own caption and retain a wider view when the reviewer needs the surrounding timeline.
Before sending, read the note and image together. The recipient should be able to identify the run, understand the observation and find the authorised supporting results. Select the reviewed export deliberately, then confirm the attachment is that file rather than the working copy beside it.
