diff --git a/CHANGELOG.md b/CHANGELOG.md
index f8bf33ff66..508b47e664 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -2,6 +2,9 @@
 
 [Full changelog](https://github.com/mozilla/glean/compare/v69.0.0...main)
 
+* General
+  * Accept 0-duration samples in timing_distribution metrics ([bug 2049040](https://bugzilla.mozilla.org/show_bug.cgi?id=2049040))
+
 # v69.0.0 (2026-06-22)
 
 [Full changelog](https://github.com/mozilla/glean/compare/v68.0.0...v69.0.0)
diff --git a/docs/user/reference/metrics/timing_distribution.md b/docs/user/reference/metrics/timing_distribution.md
index 704de4bffd..46c5728f79 100644
--- a/docs/user/reference/metrics/timing_distribution.md
+++ b/docs/user/reference/metrics/timing_distribution.md
@@ -28,7 +28,8 @@ controls the minimum and maximum values that will recorded:
   - `millisecond`: 1ms <= x <= ~19 years
 
 Overflowing this range is considered an error and is reported through the error reporting mechanism.
-Underflowing this range is not an error and the value is silently truncated to the minimum value.
+Underflowing this range is not an error and the value is silently promoted to the minimum value.
+Values of zero are preserved as zero.
 
 Additionally, when a metric comes from GeckoView (the `geckoview_datapoint` parameter is present),
 the `time_unit` parameter specifies the unit that the samples are in when passed to Glean.
diff --git a/glean-core/src/metrics/timing_distribution.rs b/glean-core/src/metrics/timing_distribution.rs
index 5d8dd9c053..7581dce130 100644
--- a/glean-core/src/metrics/timing_distribution.rs
+++ b/glean-core/src/metrics/timing_distribution.rs
@@ -241,7 +241,7 @@ impl TimingDistributionMetric {
         let min_sample_time = self.time_unit.as_nanos(1);
         let max_sample_time = self.time_unit.as_nanos(MAX_SAMPLE_TIME);
 
-        duration = if duration < min_sample_time {
+        duration = if duration < min_sample_time && duration > 0 {
             // If measurement is less than the minimum, just truncate. This is
             // not recorded as an error.
             min_sample_time
@@ -389,9 +389,7 @@ impl TimingDistributionMetric {
                     // Check the range prior to converting the incoming unit to
                     // nanoseconds, so we can compare against the constant
                     // MAX_SAMPLE_TIME.
-                    if sample == 0 {
-                        sample = 1;
-                    } else if sample > MAX_SAMPLE_TIME {
+                    if sample > MAX_SAMPLE_TIME {
                         num_too_long_samples += 1;
                         sample = MAX_SAMPLE_TIME;
                     }
@@ -502,7 +500,7 @@ impl TimingDistributionMetric {
             for &sample in samples.iter() {
                 let mut sample = sample;
 
-                if sample < min_sample_time {
+                if sample < min_sample_time && sample > 0 {
                     sample = min_sample_time;
                 } else if sample > max_sample_time {
                     num_too_long_samples += 1;
@@ -664,9 +662,7 @@ impl<'a> LocalTimingDistribution<'a> {
         // Check the range prior to converting the incoming unit to
         // nanoseconds, so we can compare against the constant
         // MAX_SAMPLE_TIME.
-        let sample = if sample == 0 {
-            1
-        } else if sample > MAX_SAMPLE_TIME {
+        let sample = if sample > MAX_SAMPLE_TIME {
             self.errors += 1;
             MAX_SAMPLE_TIME
         } else {
diff --git a/glean-core/tests/memory_distribution.rs b/glean-core/tests/memory_distribution.rs
index e81e7ecdbb..2f1faeb033 100644
--- a/glean-core/tests/memory_distribution.rs
+++ b/glean-core/tests/memory_distribution.rs
@@ -127,7 +127,7 @@ fn the_accumulate_samples_api_correctly_stores_memory_values() {
 
     // Accumulate the samples. We intentionally do not report
     // negative values to not trigger error reporting.
-    metric.accumulate_samples_sync(&glean, [1, 2, 3].to_vec());
+    metric.accumulate_samples_sync(&glean, [0, 1, 2, 3].to_vec());
 
     let snapshot = metric
         .get_value(&glean, "store1")
@@ -138,9 +138,13 @@ fn the_accumulate_samples_api_correctly_stores_memory_values() {
     // Check that we got the right sum of samples.
     assert_eq!(snapshot.sum, 6 * kb);
 
-    // We should get a sample in 3 buckets.
+    // Check that we have the right count of samples.
+    assert_eq!(snapshot.count, 4);
+
+    // We should get a sample in 4 buckets.
     // These numbers are a bit magic, but they correspond to
     // `hist.sample_to_bucket_minimum(i * kb)` for `i = 1..=3`.
+    assert_eq!(1, snapshot.values[&0]);
     assert_eq!(1, snapshot.values[&1023]);
     assert_eq!(1, snapshot.values[&2047]);
     assert_eq!(1, snapshot.values[&3024]);
diff --git a/glean-core/tests/timing_distribution.rs b/glean-core/tests/timing_distribution.rs
index cd85f62fe7..4cbfa342c0 100644
--- a/glean-core/tests/timing_distribution.rs
+++ b/glean-core/tests/timing_distribution.rs
@@ -167,7 +167,7 @@ fn the_accumulate_samples_api_correctly_stores_timing_values() {
 
     // Accumulate the samples. We intentionally do not report
     // negative values to not trigger error reporting.
-    metric.accumulate_samples_sync(&glean, &[1, 2, 3]);
+    metric.accumulate_samples_sync(&glean, &[0, 1, 2, 3]);
 
     let snapshot = metric
         .get_value(&glean, "store1")
@@ -179,11 +179,12 @@ fn the_accumulate_samples_api_correctly_stores_timing_values() {
     assert_eq!(snapshot.sum, 6 * seconds_to_nanos);
 
     // Check that we got the right number of samples.
-    assert_eq!(snapshot.count, 3);
+    assert_eq!(snapshot.count, 4);
 
-    // We should get a sample in 3 buckets.
+    // We should get a sample in 4 buckets.
     // These numbers are a bit magic, but they correspond to
     // `hist.sample_to_bucket_minimum(i * seconds_to_nanos)` for `i = 1..=3`.
+    assert_eq!(1, snapshot.values[&0]);
     assert_eq!(1, snapshot.values[&984625593]);
     assert_eq!(1, snapshot.values[&1969251187]);
     assert_eq!(1, snapshot.values[&2784941737]);
@@ -399,7 +400,7 @@ fn raw_samples_api_error_cases() {
     metric.accumulate_raw_samples_nanos_sync(
         &glean,
         &[
-            0,                   /* rounded up to 1 */
+            0,                   /* valid */
             1,                   /* valid */
             max_sample_time + 1, /* larger then the maximum, will record an error and the maximum */
         ],
@@ -410,7 +411,7 @@ fn raw_samples_api_error_cases() {
         .expect("Value should be stored");
 
     // Check that we got the right sum.
-    assert_eq!(snapshot.sum, 2 + max_sample_time as i64);
+    assert_eq!(snapshot.sum, 1 + max_sample_time as i64);
 
     // Check that we got the right number of samples.
     assert_eq!(snapshot.count, 3);
@@ -418,7 +419,8 @@ fn raw_samples_api_error_cases() {
     // We should get a sample in 3 buckets.
     // These numbers are a bit magic, but they correspond to
     // `hist.sample_to_bucket_minimum(i * seconds_to_nanos)` for `i = {1, max_sample_time}`.
-    assert_eq!(2, snapshot.values[&1]);
+    assert_eq!(1, snapshot.values[&0]);
+    assert_eq!(1, snapshot.values[&1]);
     assert_eq!(1, snapshot.values[&599512966122]);
 
     // 1 error should be reported.