# Dumping CommitDebug information when latency spike?

**URL:** <https://forums.foundationdb.org/t/dumping-commitdebug-information-when-latency-spike/5021>\
**Category:** Using FoundationDB\
**Created:** [August 14, 2025, 8:39pm UTC](https://forums.foundationdb.org/t/dumping-commitdebug-information-when-latency-spike/5021 "2025-08-14T20:39:17Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![mpatou\_openai](https://sea1.discourse-cdn.com/foundationdb/user_avatar/forums.foundationdb.org/mpatou_openai/32/1883_2.png) [@mpatou\_openai](https://forums.foundationdb.org/u/mpatou_openai)\
**Post date:** [August 14, 2025, 8:39pm UTC](https://forums.foundationdb.org/t/dumping-commitdebug-information-when-latency-spike/5021/1 "2025-08-14T20:39:17Z")

</div>

We are facing situation when we have sudden spikes in the p99.9 latency of commits and we would like to investigate a bit more. I looked at CommitDebug information roughly doing the delta of the timestamp of message with location `CommitProxyServer.commitBatch.AfterReportRawCommittedVersion` which should be more or less the last message - message with location `CommitProxyServer.commitBatch.Before` but it seems because of sampling I’m only getting message that somewhat show the average value maybe p90 once in a while.  
I’m wondering if there is a way to trigger the generation of message if the overall latency of the transaction was \>Xms