We are seeing a v1.9 gateway go into a reload loop the moment we deploy our dashboard for v2.3 gateway, i.e a v1.3 dashboard for a v2.3 gateway. Both the v1.9 and v2.3 deployments are using a single redis cluster, but different redis DBs in that cluster. There is no other configuration common between the 2 version deployments(no common shared node value, admin secret, node secret etc). How can deploying a v1.3 dashboard trigger a reload loop in an altogether different gateway which is running on v1.9? Any help would be appreciated. Thanks in advance.
Dashboard v1.3 isn’t compatible with anything but v2.3, the two aren’t designed to be mixed and matched. It causes a reload loop 
You’ll need v2.3 with v1.3, the reload loop is caused by the gateway interpreting the heartbeat messages sent by dash as reload signals.
Martin, these are 2 different environments. The v1.9 gateway is a stage deployment of tyk and we are running another test deployment with v2.3 gw and v1.3 dashboard. However, the test dashboard(v1.3) is causing the stage gw(v1.9) to go into a loop just after the dashboard comes up. What is the mechanism for heartbeat from dashboard to go to the gateway, is it directly or does it set something on redis while sending the heartbeat ? We have a common redis cluster(with different DBs in the cluster) for these 2 environments.
Ah, I see - If the environments are sharing a redis DB, then the v1.9 gateway will hot reload loop because it will be getting the signals from the stage dash.