Hello Trickster team,
Currently, Trickster provides an endpoint ( http://8484/trcikster/config/reload ) for configs reloading and this is working good.
In our environment, we run Trickster in K8s with 10 (may be more in future) pods deployment. We have implemented a helper app that does the heavy lifting of updating configmaps (for example, a new member joining an ALB backend) and invoke the reload endpoint at every pod.
This supposed to be a simple task, however, in reality, there are failing factors within K8s runtime. For example, pods may not come up at the same time when helper pokes the kube-api.
Ideally, if Trickster can watch for configs (file contents) change (at file system level) and call reload internally, it will be a great help at our side not to worry if any pods are missing. Also this is a more "native" approach than getting an external app to do it one by one.
Thanks,
Keith
Hello Trickster team,
Currently, Trickster provides an endpoint ( http://8484/trcikster/config/reload ) for configs reloading and this is working good.
In our environment, we run Trickster in K8s with 10 (may be more in future) pods deployment. We have implemented a helper app that does the heavy lifting of updating configmaps (for example, a new member joining an ALB backend) and invoke the reload endpoint at every pod.
This supposed to be a simple task, however, in reality, there are failing factors within K8s runtime. For example, pods may not come up at the same time when helper pokes the kube-api.
Ideally, if Trickster can watch for configs (file contents) change (at file system level) and call reload internally, it will be a great help at our side not to worry if any pods are missing. Also this is a more "native" approach than getting an external app to do it one by one.
Thanks,
Keith