diff --git a/BedrockServer.cpp b/BedrockServer.cpp
index c0cad65f3..50420134f 100644
--- a/BedrockServer.cpp
+++ b/BedrockServer.cpp
@@ -1030,8 +1030,6 @@ void BedrockServer::runCommand(unique_ptr<BedrockCommand>&& _command, bool isBlo
                             // loop and send it to followers. NOTE: we don't check for null here, that should be
                             // impossible inside a worker thread.
                             _syncNode->notifyCommit();
-                            SINFO("Committed leader transaction #" << transactionID << "(" << transactionHash << "). Command: '" << command->request.methodLine << "', blocking: "
-                                  << (isBlocking ? "true" : "false"));
                             _conflictManager.recordTables(command->request.methodLine, db.getTablesUsed());
                             // So we must still be leading, and at this point our commit has succeeded, let's
                             // mark it as complete. We add the currentCommit count here as well.
diff --git a/libstuff/libstuff.cpp b/libstuff/libstuff.cpp
index 4f8370a69..9fcdaaf0d 100644
--- a/libstuff/libstuff.cpp
+++ b/libstuff/libstuff.cpp
@@ -2722,7 +2722,6 @@ int SQuery(sqlite3* db, const char* e, const string& sql, SQResult& result, int6
 
     // But we log for commit conflicts as well, to keep track of how often this happens with this experimental feature.
     if (extErr == SQLITE_BUSY_SNAPSHOT) {
-        SHMMM("[concurrent] commit conflict.");
         return extErr;
     }
     return error;
diff --git a/sqlitecluster/SQLite.cpp b/sqlitecluster/SQLite.cpp
index 8d10d6ee1..2ef57a934 100644
--- a/sqlitecluster/SQLite.cpp
+++ b/sqlitecluster/SQLite.cpp
@@ -663,7 +663,7 @@ bool SQLite::prepare(uint64_t* transactionID, string* transactionhash) {
     SQResult journalLookupResult;
     SASSERT(!SQuery(_db, "getting commit min", "SELECT MIN(id) FROM " + _journalName, journalLookupResult));
     uint64_t minJournalEntry = journalLookupResult.size() ? SToUInt64(journalLookupResult[0][0]) : 0;
-    
+
     // Note that this can change before we hold the lock on _sharedData.commitLock, but it doesn't matter yet, as we're only
     // using it to truncate the journal. We'll reset this value once we acquire that lock.
     uint64_t commitCount = _sharedData.commitCount;
@@ -682,8 +682,6 @@ bool SQLite::prepare(uint64_t* transactionID, string* transactionhash) {
         string query = "DELETE FROM " + _journalName + " WHERE id < " + SQ(oldestCommitToKeep) + " LIMIT " + SQ(deleteLimit);
         SASSERT(!SQuery(_db, "Deleting oldest journal rows", query));
         size_t deletedCount = sqlite3_changes(_db);
-        SINFO("Removed " << deletedCount << " rows from journal " << _journalName << ", oldestToKeep: " << oldestCommitToKeep << ", count:"
-               << commitCount << ", limit: " << _maxJournalSize << ", in " << (STimeNow() - startUS) << "us.");
     }
 
     // We lock this here, so that we can guarantee the order in which commits show up in the database.
@@ -771,9 +769,6 @@ int SQLite::commit(const string& description, function<void()>* preCheckpointCal
     result = SQuery(_db, "committing db transaction", "COMMIT");
     _lastConflictPage = _conflictPage;
     _lastConflictTable = _conflictTable;
-    if (_lastConflictPage) {
-        SINFO(format("part of last conflict page: {}, conflict table: {}",  _conflictPage, _conflictTable));
-    }
 
     // If there were conflicting commits, will return SQLITE_BUSY_SNAPSHOT
     SASSERT(result == SQLITE_OK || result == SQLITE_BUSY_SNAPSHOT);
@@ -834,7 +829,7 @@ int SQLite::commit(const string& description, function<void()>* preCheckpointCal
         _lastConflictPage = 0;
         _lastConflictTable = "";
     } else {
-        SINFO("Commit failed, waiting for rollback.");
+        // The commit failed, we will rollback.
     }
 
     // if we got SQLITE_BUSY_SNAPSHOT, then we're *still* holding commitLock, and it will need to be unlocked by
@@ -883,9 +878,6 @@ void SQLite::rollback() {
         // Finally done with this.
         _insideTransaction = false;
         _uncommittedHash.clear();
-        if (_uncommittedQuery.size()) {
-            SINFO("Rollback successful.");
-        }
         _uncommittedQuery.clear();
 
         // Only unlock the mutex if we've previously locked it. We can call `rollback` to cancel a transaction without