You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
dead_letter_queue.flush_check_interval new config for flushing staled segment files. (#19036)
* Validates to be min 1s to keep consistency with the docs. Introduces new config for flushing staled segment files.
* Add pipeline name to the DLQ flush thread name for better visibility in the threads API results. Add suggestions from the docs review. Re-organize the duration clam logic in a way for better maintainable and fix the unit tests.
* Update logstash-core/src/main/java/org/logstash/common/io/DeadLetterQueueWriter.java
Remove unused method.
Co-authored-by: Andrea Selva <selva.andre@gmail.com>
* Move the flush chech interval to the DeadLetterQueueWriter.Builder. Remove confusing scheduler from the docs explanations. unit tests for the only newly introduced conditions.
* Apply suggestions from code review
Doc consistency and test rename suggestions accepted.
Co-authored-by: Andrea Selva <selva.andre@gmail.com>
* Keep the interval type as a Duration, rename and simplify test suites.
---------
Co-authored-by: Andrea Selva <selva.andre@gmail.com>
Copy file name to clipboardExpand all lines: docs/reference/dead-letter-queues.md
+7-1Lines changed: 7 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -78,8 +78,14 @@ Note that this value cannot be set to lower than 1000ms.
78
78
dead_letter_queue.flush_interval: 5000
79
79
```
80
80
81
+
Stale segments files are periodically checked if they need to be flushed. This period is controlled by the `dead_letter_queue.flush_check_interval` setting. This setting is in milliseconds, and defaults to 1000ms. A smaller value ensures faster segment rotation when infrequent writes occur, at the cost of CPU consumption with more frequent segment checks execution. A larger value reduces checks overhead but delays segment sealing, for the worst case it will be `dead_letter_queue.flush_interval` + `dead_letter_queue.flush_check_interval`. This value cannot be set to lower than 1000ms.
82
+
83
+
```yaml
84
+
dead_letter_queue.flush_check_interval: 1000
85
+
```
86
+
81
87
::::{note}
82
-
You may not use the same `dead_letter_queue` path for two different Logstash instances.
88
+
You cannot use the same `dead_letter_queue` path for two different Logstash instances.
0 commit comments