I would like to know if API gateway can load balance across servers serving up an API it is protecting, or if it needs to integrate with VIPs for load balancing / failover?
Thanks in advance.
The HTTP Routing assertion has several load balancer/fail over strategies such as round robin, order sticky, etc ( Route via HTTP(S) Assertion - CA API Gateway - 9.1 - CA Technologies Documentation ) which can also be used in conjunction with Routing Strategies (Working With Dynamic Routing - CA API Gateway - 9.1 - CA Technologies Documentation ).
The Gateway can also be used with a load balancer VIP that manages available back-end systems.
Director, CA Support
My Requirement is for API Gateway to act as a proxy to a service deployed on 2 back end servers, TEST A and TEST B.
API Gateway should balance send the request equally across both servers and send to Test B only if Test A is down and vice versa.
I am following the document you mentioned here. I tried to create a routing strategy, I selected on create New ROute, I gave it a name and clicked 'ADD' for route properties. I am confused as to what needs to go here, appreciate assistance. It may be very intuitive and self explanatory but I am stuck as the document says just '
Ravichandran, I m not trying to hijack your thread , just trying to keep related info in one place. Please let me know if you want me to start another thread.
Thank you for your help. Really useful information, addresses my query.
You should be able to do this in the Route via HTTP Assertion itself.
Here's an example of what a route could look like
it this case, we have multiple urls to route to and we are using a 'Round-Robin' failover strategy...
looking at the doc, here are the details of 'Round-Robin'
hope this helps!
I know it's the old thread but still trying my luck out. My questions is, how does the API gateway knows if the end application is does? Does it do the polling like a Industry Standard load balancer or it does it in the real time when the transaction comes?
Question appear trivial but in our case we are trying to remove the load balancer between APIG and backend application instances. So all I want to know how APIG handles this scenario.
A typical LB would do a health check to determine if the service is available or not. This happens in background and not at the time transactions comes in. We are not worried that API gateway would do the fail over but rather it would try to direct every alternate request to the unavailable server and then redirect it to the next one active. Hope I am able to clearly articulate my query.
Thank You , Sir. Much appreciated. Will test and let you know.
I have been playing with that assertion for a couple of days but could not find that screen,
We don't do polling by default within the gateway but you could look to use Dynamic Routing (Working With Dynamic Routing - CA API Gateway - 9.2 - CA Technologies Documentation) in combination with using Scheduled tasks that could ping the back end at periodic intervals and update the information associated with Dynamic Routing.
thanks for this input stephen , it was very helpful for s similar use case for me.