We use a InfluxDB to store the graphite metrics generated by Gatling. We are experiencing a connection loss when the InfluxDB is restarted. The connection is not restored when the InfluxDB is up and running again. The InfluxDB is running in a kubernetes cluster and the Gatling load generator is running on a Wintel server environment VM.
How to prevent and or restore this connectivity during a running test.
There’s an ongoing contribution to revisit the Graphite TcpSender: Graphite: Revised TcpSender by vJoeyz · Pull Request #4202 · gatling/gatling · GitHub
Sadly, it’s currently broken. Feel free to contribute.