diff --git a/gnocchi/README.md b/gnocchi/README.md
index 6fb7bd0..ea63666 100644
--- a/gnocchi/README.md
+++ b/gnocchi/README.md
@@ -1,5 +1,6 @@
-# Gnocchi
+## Gnocchi
 
+# blabla commercial
 
 https://gnocchi.xyz/
 
diff --git a/influxdb/README.md b/influxdb/README.md
index 97400cb..ce37442 100644
--- a/influxdb/README.md
+++ b/influxdb/README.md
@@ -1,7 +1,8 @@
 
 
-# InfluxDB
+## InfluxDB
 
+# Blabla commercial
 
 _Time Series Databases have to deal with specific workloads and requirements.
 They need to ingest millions of data points per second; to perform real-time
@@ -9,4 +10,11 @@ queries across these large data sets in a non-blocking manner; to downsample
 and evict high-precision low-value data; to optimize data storage to reduce
 storage costs; and to perform complex time-bound queries to extract meaningful
 insight from the data. It is possible to meet these requirements only with a
-purpose-built platform that InfluxData provides._
\ No newline at end of file
+purpose-built platform that InfluxData provides._
+
+# On essaye ?
+
+Ok, c'est parti. On va écrire un injecteur en Perl.
+
+
+
diff --git a/influxdb/injecteur.pl b/influxdb/injecteur.pl
new file mode 100755
index 0000000..7fd0401
--- /dev/null
+++ b/influxdb/injecteur.pl
@@ -0,0 +1,5 @@
+#!/usr/bin/perl
+
+use strict;
+
+print "injecteur v 0\n";