diff --git a/src/content/docs/ci-insights/flaky-test-detection.mdx b/src/content/docs/ci-insights/flaky-test-detection.mdx
index fc78e30ed7..1c51509106 100644
--- a/src/content/docs/ci-insights/flaky-test-detection.mdx
+++ b/src/content/docs/ci-insights/flaky-test-detection.mdx
@@ -37,7 +37,7 @@ digraph {
   subgraph cluster_commit1 {
     class="batch";
     label="Commit abc123";
-    commit1 [label="test_something2", shape=oval, class="config"];
+    commit1 [label="test_something", shape=oval, class="config"];
   }
   run1 [label="Run #1 — tests pass", class="merged"];
   run2 [label="Run #2 — tests fail", class="failed"];
diff --git a/src/content/docs/ci-insights/jobs.mdx b/src/content/docs/ci-insights/jobs.mdx
index 25bc3e4bc3..418c182f8c 100644
--- a/src/content/docs/ci-insights/jobs.mdx
+++ b/src/content/docs/ci-insights/jobs.mdx
@@ -50,3 +50,10 @@ and related test results.
   duration to see whether the failure correlates with a performance
   change.
 :::
+
+## Data Retention
+
+CI Insights keeps 30 days of job history. The free tier keeps only the last 24
+hours, so a date range reaching further back returns nothing older than that
+rather than an error. Both views show a banner when your plan is on the shorter
+window.