diff --git a/pkg/checksum/checker.go b/pkg/checksum/checker.go
index cce4b14d..bcd2477c 100644
--- a/pkg/checksum/checker.go
+++ b/pkg/checksum/checker.go
@@ -21,8 +21,6 @@ import (
 	"golang.org/x/sync/errgroup"
 )
 
-var binlogFlushIntervalDuringChecksum = 30 * time.Second
-
 type Checker struct {
 	sync.Mutex
 	table       *table.TableInfo
@@ -209,7 +207,7 @@ func (c *Checker) Run(ctx context.Context) error {
 	// - If they are there, they will take a huge amount of time to flush
 	// - The memory requirements for 1MM deltas seems reasonable, but for a multi-day
 	//   checksum it is reasonable to assume it may exceed this.
-	go c.feed.StartPeriodicFlush(ctx, binlogFlushIntervalDuringChecksum)
+	go c.feed.StartPeriodicFlush(ctx, repl.DefaultFlushInterval)
 	defer c.feed.StopPeriodicFlush()
 
 	g, errGrpCtx := errgroup.WithContext(ctx)
diff --git a/pkg/migration/runner.go b/pkg/migration/runner.go
index 3dc38514..a6718a4b 100644
--- a/pkg/migration/runner.go
+++ b/pkg/migration/runner.go
@@ -43,10 +43,6 @@ var (
 	checkpointDumpInterval  = 50 * time.Second
 	tableStatUpdateInterval = 5 * time.Minute
 	statusInterval          = 30 * time.Second
-	// binlogPerodicFlushInterval is the time that the client will flush all binlog changes to disk.
-	// Longer values require more memory, but permit more merging.
-	// I expect we will change this to 1hr-24hr in the future.
-	binlogPerodicFlushInterval = 30 * time.Second
 )
 
 func (s migrationState) String() string {
@@ -429,7 +425,7 @@ func (r *Runner) setup(ctx context.Context) error {
 	// Continuously update the min/max and estimated rows
 	// and to flush the binary log position periodically.
 	go r.table.AutoUpdateStatistics(ctx, tableStatUpdateInterval, r.logger)
-	go r.replClient.StartPeriodicFlush(ctx, binlogPerodicFlushInterval)
+	go r.replClient.StartPeriodicFlush(ctx, repl.DefaultFlushInterval)
 	return nil
 }
 
diff --git a/pkg/repl/client.go b/pkg/repl/client.go
index d3f741af..d9d41264 100644
--- a/pkg/repl/client.go
+++ b/pkg/repl/client.go
@@ -33,6 +33,10 @@ const (
 	// multiple-flush-threads, which should help it group commit and still be fast.
 	DefaultBatchSize = 1000
 	flushThreads     = 16
+	// DefaultFlushInterval is the time that the client will flush all binlog changes to disk.
+	// Longer values require more memory, but permit more merging.
+	// I expect we will change this to 1hr-24hr in the future.
+	DefaultFlushInterval = 30 * time.Second
 )
 
 type queuedChange struct {