Replies: 2 comments 4 replies
-
hey @kebe7jun can you share the specific steps to repro this issue |
Beta Was this translation helpful? Give feedback.
1 reply
-
looks like related to |
Beta Was this translation helpful? Give feedback.
3 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
We are currently working on EnvoyGateway in hopes of bring it into our product and have encountered this issue.
Scenario: We will deploy a public EnvoyGateway instance for a cluster to expose the services in the cluster to the outside world. Here, we may design Layer 4 and Layer 7 traffic at the same time.
Problem: For layer 4 traffic, such as SSH ports, we may often add/remove instances on the backend that provide SSH services. In these cases, we need to configure Gateway and TCPRoute to ensure that traffic can reach the actual Pod from the user side. But once I modify the Gateway's
listeners
configuration, the Deployment of the Gateway will be restarted, which is unacceptable for a public gateway component. However, we cannot create a new Gateway instance for each port/instance, which will greatly increase our operating costs.This may be caused by re-rendering Container's
ports
? But in fact this field is not used in the normal traffic path (Service's targetPort is a directly defined port, not a name). Maybe there are other reasons, but I didn't delve into it.So, can EnvoyGateway solve this problem? Or provide an optional parameter for the user to decide? I've looked through the documentation and can't seem to find a similar configuration.
AFAIK, some other similar gateway implementations do not seem to have similar problems.
Beta Was this translation helpful? Give feedback.
All reactions