johannes-engler-mw commented on issue #2514:
URL:
https://github.com/apache/apisix-ingress-controller/issues/2514#issuecomment-3575686405
Hi @juzhiyuan,
We are evaluating the API-driven standalone mode for use with the APISIX
Ingress Controller in our Kubernetes environment. The goal is to simplify our
architecture by removing the dependency on etcd.
However, based on the current documentation and our research, we think that
we identified several limitations that make it difficult to adopt this mode
outside of testing. We would like to ask about the roadmap to General
Availability (GA) for this feature and whether the path to GA includes fixes
for the following issues:
Key Concerns:
- Full Configuration Push on Updates: The current implementation requires
the Ingress Controller to push the entire configuration for every single
change. This creates a significant scalability bottleneck as our route count
grows. Will the GA version support incremental updates (e.g., only updating the
changed route/upstream) to improve performance?
- No Persistence & Recovery Downtime: Since APISIX starts with an empty
configuration in this mode, any pod restart results in downtime until the
Ingress Controller can re-sync. Are there plans to introduce a persistence or
caching mechanism (like writing to a local file) that would allow APISIX to
recover instantly on its own?
- Lack of Dashboard Support: The APISIX Dashboard does not work without
etcd, which removes a critical tool for observability and troubleshooting. Will
there be an alternative way to visualize the running configuration in
standalone mode?
- Will achieving full feature parity with the etcd mode be a requirement for
GA?
Thank you for your excellent work on this project!
BR,
Johannes
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]