I'm trying to take one request and route it to multiple backend targets.
I'm NOT trying to take one request and route it to one backend target via multiple rules as a lot of the App Gateway documentation suggests.
What I've tried:
I've tried creating 2 of the same listener's (which isn't possible) so that I can create 2 different rules targeting different backend pools for each listener.
I set up one rule with multiple backend target's but I'm unable to have the same path with different backend targets. I tried specifying a wildcard path and an explicit path but would get an HTTP error when making requests through App Gateway.
Is it possible to duplicate a request and send each copy to a different backend target?
Not sure if this will work. I think you need a different pool for each listener.
Can you try creating a separate pool for each of your backends?
Then create a listener in that pool.
Then create a rule to do the forwarding.
Ian
Related
For a Flutter web app, I'm using a package that creates its own HTTP client instance, communicating with some gateway.
Additionally I have my own server for the app.
Should I modify the package, to have both connections on the same instance?
Or should I have a HTTP client for every connection?
The benefit of reusing an instance of HttpClient is that Dart can then let a connection to stay open for later reuse in case you are going to make another request against the same server.
So I would in general recommend reusing a HttpClient instance if possible, but here are no issue in having multiple instances of it. Especially if the two instances are used to connect to different servers.
There can be several reasons for having multiple instances like e.g. you want different connection settings (e.g. timeout, user-agent, different handling of certificates) for different endpoints.
While trying to create a web application for shared drawing I got stuck on a problem regarding Kubernetes and scaling. The application uses an ASP.NET Core backend with SignalR for sharing the drawing data across its users. For scaling out the application I am using a deployment for each microservice of the system. For the SignalR part though, additional configuration is required.
After some research I have found out about the possibility to sync all instances of the SignalR backend either through the use of Azures SignalR Service or the use of a Redis backplane. The latter of which I have gotten to work on my local minikube environment. I am not really happy with this solution because of the following reasons:
My main concern is that like this I have created a hard bottleneck in
the system. Unlike in a chat application where data is sent only once
in a while, messages are sent for every few points drawn in the
shared drawing experience by any client. Simply put, a lot of traffic
can occur and all of it has to pass through the single Redis backplane.
Additionally to me it seems unneccessary to make all instances of the SignalR backend talk to each
other. In this application shared drawing does only occur in small groups of up to 10 clients lets
say. Groups of this size can easily be hosted on a single instance.
So without syncing all instances of the SignalR backend I would have to route the clients connection based on the SignalR group name to the right instance of the SignalR backend when the client is trying to join a group.
I have found out about StatefulSets which allow me to have a persistent address for each backend pod in the cluster. I then could somehow associate the SignalR group IDs with the pod addresses they are running on in lets say another look up microservice. The problem with this is that the client needs to be able to access the right pod from outside of the cluster where that cluster internal address does not really help.
Also I am wondering if there isnt a whole better approach to the problem since I am very new to the world of kubernetes. I would be very greatful for your thoughts on this issue and any hint towards a (better) solution.
I have two apps I wanted to have "fully managed" by Cloud Run. One is a pure Vue.js SPA and the other is the belonging backend server for it that is connected to a MySQL and also fetches some other API endpoints.
Now I have deployed both apps but am totally unaware on how I can give the frontend app access to the backend app. They should be both running on the same domain to avoid the frontend from.
Current URL of the frontend app: https://myapp-xl23p3zuiq-ew.a.run.app
So I'd love to have the server accessible by: https://myapp-xl23p3zuiq-ew.a.run.app/api
Is this somewhat possible to achieve with Cloud Run?
I was having the same issue. The general idea that one usually has is to use path mapping and map / to your client and /server to your backend. After googling for a while I found this:
https://cloud.google.com/run/docs/mapping-custom-domains
Base path mapping: not supported
The term base path refers to the URL
path name that is after the domain name. For example, users is the
base path of example.com/users. Cloud Run only allows you to map a
domain to /, not to a specific base path. So any path routing has to
be handled by using a router inside the service's container or by
using Firebase Hosting.
Option1:
I ended up creating an "all in one" docker image with an nginx as reverse proxy and the client (some static files) and server (in my case a python application powered by uwsgi).
If you are looking for inspiration, you can check out the public repository here: https://gitlab.com/psono/psono-combo
Opttion2:
An alternative would be to host your client on client.example.com, your server on server.example.com and then create a third docker run instance with a reverse proxy under example.com.
All requestes would be "proxied" to the client and server. Your users will only interact with example.com so CORS won't be an issue.
Option3:
Configure CORS, so people accessing example.com can also connect to server.example.com
Currently this is not possible in Cloud Run, as already said on the comments to your question.
You could check if there are any Feature Request for this functionality on Buganizer (Google Issue Tracker), currently there seems to be none, and if that is indeed the case, you can create a new Feature Request by changing the request type from Bug to Feature Request and as Google develops it on their road map, you will be informed.
Hope this helped you.
I have some legacy application, which deployed on clustered environment. When one of the application nodes receives call it gets from some configuration file static list of all application nodes where application is deployed.
When all ip's collected it communicates with each app node over jmx.
Current aim is to migrate to k8s, so in this case list of application pods is dynamic and can be just stored as is. Need to implement something like service discovery.
Current thoughts is to implement some simple rest service that will run in separate pod, main aim of which is always return some list of ips (entrypoints) of application pods filtered by some predicate.
So I have few questions:
Is it correct way to work? Any other options? (without changing legacy code)
Is there any ready solution for this? If not, how can I get information about needed pods inside my rest service?
Define a service with a scope selector so all your special pods are included then you can list all your endpoints IP's asking the apiservice.
You can check it's working with the command.
kubectl get endpoints
After that remains how to execute this command inside your pod. That's another story.
This link explain that matter
https://kubernetes.io/docs/tasks/administer-cluster/access-cluster-api/#accessing-the-api-from-a-pod
Looks you're running a clustered application, so probably you need a Headless Service combined with a StatefulSet.
With this, you will be able to reach your replicas using simple DNS like replicas-[0-9].namespace.svc without need to extract IP addresses from endpoints query.
I have two web roles scaled up to 4 instances and at least two instances are running at a time. The certificate requested by a user via first website is pushed to blob storage and creates a queue message for the other website to read it.
I have a while loop inside the Run method implementing RoleEntryPoint to check the message, upload the certificate to store if available, apply https binding, and delete the message.
The bindings are applied to current instance and not into both. Is there a way to I can access all instances and apply bindings to both at once? If not, what are some better approaches?