> ## Documentation Index > Fetch the complete documentation index at: https://docs.peepsai.com/llms.txt > Use this file to discover all available pages before exploring further. # Find what failed > Use a release report's three tabs: Issues for the failures grouped by cause, Where it fails for where they concentrate, and All results for every latest result. A [release report](https://docs.peepsai.com/releases/read-a-report.md) has three tabs. Each reads the same results in a different way. | Tab | Use it to | | - | - | | **Issues** | Work through the failures, grouped into issues. | | **Where it fails** | See which areas and configurations the failures concentrate in. | | **All results** | Look up any test's latest result on any configuration, passing ones included. | ## Issues An issue is a group of failures that look like one problem. The number on its left is how many failing results it holds, and the strip on its right shows them by configuration. ![The Issues tab: four cards counting failures by Analysis pending, Possible product issue, Analysis failed and Possible environment issue, then a search box, a Hide reviewed switch and an open issue with Mark reviewed and Not a bug buttons, an AI analysis marked Likely cause, unconfirmed, and its errors grouped by similar message.](https://mintcdn.com/peeps-ai/dn4jpKrOiHjTnw5R/images/releases/issues-tab.png?fit=max\&auto=format\&n=dn4jpKrOiHjTnw5R\&q=85\&s=efe17539fea014bd9572a0fe1985bf77) Select an issue to open it in the list: * **Mark reviewed** and **Not a bug** record your review. Select the active one again to clear it. Contributors and above can do this. A review changes no result. * **AI analysis** gives a likely cause, always marked unconfirmed, with a confidence. **Copy for AI** copies the issue for your coding agent. * **Errors** groups its results by similar message. Open one to reach the results and their runs. * **By configuration** lists how many failing results each configuration has. * **Open issue** opens the issue's own page. A row's second line shows its likely cause or analysis state, how many configurations it fails on, and its history: * **Since** names the earliest release Peeps has a record of it in. * **No earlier record** means Peeps found none. It may only mean Peeps has no history for it. ### Cards The cards above the list count failures and issues by likely cause, or by the configuration, area or ticket when you [list issues that way](https://docs.peepsai.com/releases/filter-and-share.md#group). Select a card to narrow the list, and select more to add to it. A cause can be an unconfirmed AI label, so narrowing by one changes no outcome. ### Tracked issues When your organization has connected Linear or Jira, each issue shows where it stands in your tracker, such as **Not tracked** or **Done, failed since**. Contributors can create a ticket or link an existing one from the row, or from the issue's own page. See [Issue trackers](https://docs.peepsai.com/mcp/issue-trackers.md). The ticket goes on the issue itself, with the row's run linked beside it, so it shows on that issue in every release the issue fails in, and on no other issue. The ticket's web link in your tracker opens the issue's page. When the issue was merged into another, the ticket goes on the issue it was merged into. To take a ticket off an issue, use **Unlink** on the issue's page across releases. It removes the ticket from that issue and its runs only; the same ticket on another issue stays. ### Search and hide **Search issues** looks through the list. **Hide reviewed** hides what you've reviewed; it applies to you alone. ## Where it fails This tab shows the failures as a grid, with areas down the side and configurations, or the dimension you [group by](https://docs.peepsai.com/releases/filter-and-share.md#group), across the top. ![The Where it fails tab with an Issues seen before chip and 7 failing results in other issues. Across the top are bars and failure rates for Chrome, Safari and Firefox. The Checkout area reads 3 of 18 cases with red cells, one case row is marked Filtered out with an empty outlined cell, and the Accounts / Sign-in area reads No failures match these filters with outlined cells.](https://mintcdn.com/peeps-ai/dn4jpKrOiHjTnw5R/images/releases/where-it-fails.png?fit=max\&auto=format\&n=dn4jpKrOiHjTnw5R\&q=85\&s=7ee6b4bedcccdaf68f53d5ff3e4bdfea) * **The heading** shows failing results, how many results were decided (passed, failed or errored), and the failure rate, per column. Select a column's heading to filter the whole page to it. * **An area row** counts the test cases with a failure, of those that ran. Open it to see its test cases, whose cells count failing results. * **A cell** opens the failing results behind it. | Cell | Means | | - | - | | Red, darker with a larger share failing | Holds failures. | | Grey | Ran with no failure here. | | Empty | Didn't run here. | | Outlined, empty | Holds failures, all in issues your [selectors](https://docs.peepsai.com/releases/filter-and-share.md#issue-selectors) leave out. **Filtered out** isn't a pass. | With a selector on, areas, test cases and cells count only the issues you selected, and an area with nothing left says **No failures match these filters**. The columns' totals keep counting every failure. A cell counts test cases while the list it opens counts results, so one test case failing on two variants reads 1 and lists 2. ## All results **All results** lists each test case's latest result per configuration, with its state, test key and run. ![The All results tab with an Issues seen before chip and 7 rows hidden, three cards for Failing, Passed and Skipped, a search box, and rows for Chrome showing each test case, its state, its test key and a Run link.](https://mintcdn.com/peeps-ai/dn4jpKrOiHjTnw5R/images/releases/all-results.png?fit=max\&auto=format\&n=dn4jpKrOiHjTnw5R\&q=85\&s=462c88c5a39eb6c63bc0e545c15a9b83) * The cards count results by state. Select one to filter the page to that state. * **Search** looks through test cases, configurations and test keys. * Select a row to open it. With a selector on, a row shows if it holds a failure in an issue you selected, or no failure at all. A row that shows keeps all its results, so a selected and an unselected failure on one row still read **Failing**. ## Move through a long list **Issues** and **All results** show 25 rows a page. **Rows per page** changes that, and the pager at the foot moves between pages. Your page is kept in the link. Changing a selector goes back to the first page of **Issues**. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.