You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Description:
AFAIK there is no proper mechanism to shutdown the Siddhi runner gracefully. I deployed the Siddhi runner in a K8s cluster. I had a requirement to enable zero downtime to the Siddhi runner deployment in K8s. I deployed the Siddhi app with HTTP source and route traffic using ingress NGINX controller.
Sometimes Siddhi runner terminates and another Siddhi runner deployment starting to consume HTTP requests. The terminating Siddhi runner has already established HTTP connections. But unfortunately, Siddhi runner shut down immediately without a response to the established connections. That leads to a 502 bad gateway in client-side.
In the termination, Siddhi runner gets a SIGTERM from K8s. At that time Siddhi runner should finish all the running threads and then starts to terminate.
Description:
AFAIK there is no proper mechanism to shutdown the Siddhi runner gracefully. I deployed the Siddhi runner in a K8s cluster. I had a requirement to enable zero downtime to the Siddhi runner deployment in K8s. I deployed the Siddhi app with HTTP source and route traffic using ingress NGINX controller.
Sometimes Siddhi runner terminates and another Siddhi runner deployment starting to consume HTTP requests. The terminating Siddhi runner has already established HTTP connections. But unfortunately, Siddhi runner shut down immediately without a response to the established connections. That leads to a 502 bad gateway in client-side.
In the termination, Siddhi runner gets a SIGTERM from K8s. At that time Siddhi runner should finish all the running threads and then starts to terminate.
Suggested Labels:
type/fix
Affected Product Version:
v5.1.0-m1
Steps to reproduce:
The text was updated successfully, but these errors were encountered: