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
Is your feature request related to a problem? Please describe.
Currently there is no option for a graceful shutdown.
When using "external" mechanisms for health checking (i.e. not relying on Readiness or Endpoints from k8s) there is no way to start failing health checks, gracefully terminate existing connections, prevent new incoming connections, etc.
Describe the solution you'd like
Some combination of scripted health check failures, listener draining, configurable terminationGracePeriodSeconds on the proxy deployment. Ideally exposed by a single helm value (?)
Describe alternatives you've considered
Some of the above mentioned solutions could be implemented after the Gloo installation
Additional context
Similar discussion in #1084
Ideally, with a graceful shutdown, ALBs relying on solely health checking envoy instances can perform e.g. an in-place upgrade with minimal-to-none downtime
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem? Please describe.
Currently there is no option for a graceful shutdown.
When using "external" mechanisms for health checking (i.e. not relying on
Readiness
orEndpoints
from k8s) there is no way to start failing health checks, gracefully terminate existing connections, prevent new incoming connections, etc.Describe the solution you'd like
Some combination of scripted health check failures, listener draining, configurable
terminationGracePeriodSeconds
on the proxy deployment. Ideally exposed by a single helm value (?)Describe alternatives you've considered
Some of the above mentioned solutions could be implemented after the Gloo installation
Additional context
Similar discussion in #1084
Ideally, with a graceful shutdown, ALBs relying on solely health checking envoy instances can perform e.g. an in-place upgrade with minimal-to-none downtime
The text was updated successfully, but these errors were encountered: