You basically set up a bunch of test URLs for your service, e.g.:
And then Tyk will ping those hosts for you with those requests and payloads, now you don;t need to always have the same host, or the same endpoint, you could set up a test for all of your services and Tyk willg enerate an uptime chart of the latency and availability for you.
To know when a host is down, you need to hook into the
HostDown event, which fires if a URL starts to fail.
If you are using the load balancer, then you can enable an option for Tyk to link the two and remove the hostname from the loadbalancer so that no more traffic hits that hostname until the uptime tet passes again.
If you want them to be linked, the hostnames must match, not the URLs
This would work
You would need to hook into the
HostDown events which can generate a webhook or run some JS.