Between the client and the discovery server, there will be another service which will redirect clients' calls to the discovery server. This service usually called Gateway service.
Hi,I think you should do these kinds of workshops more often with the students in person rather than creating youtube videos, this way we can get to know the questions and answers asked by the students in the video itself, and I must say some of the questions were very spot on, which clarified things furthur. If you make a youtube video, we need to look at the comments to see if there is a questions/clarifications asked anywhere which is very difficult since there are many appreciation comments in between.
Great course and teaching. But I think you didn't explain the server side discovery pattern properly. The server side involves a load balancer in between and every request goes through it. And few other sites say it involves more hops.
He didn't take into consideration the response Hops which seem to be the issue in the comments. And Also the discover of the discovery server, which I believe will be picked up in future videos. But other than that I felt the video was really good and on point in conveying exactly what was needed and not muddling one with unnecessary concepts. Really Thank you for this series.
Can the discovery server handle load balancing as well? For example, if the client wants to talk to service 2(there are more than 1 instance of it) can discovery server balance load among them.
@Java Brains, @Koushik - let's say I have one microservices which is build in another technology apart from springboot and I want to register that to springboot Eureka server ? Can we do that? Assume that microservices is in NODE ?
I have one question .. what if Discovery client goes down? do we have any mechanism recover that. Bcz I find client side Discovery risky bcz everything than depends on one hope 🤨
How are there no extra hops? There whould be no extra hops if there were no data to retun to the client. On the server side case we got: Client -> Discovery -> Mircroservice -> Discovery ->Client On the client case we got : Client -> Discovery -> Client-> Microservice -> Client 4 hops in each case except if on the 1st case when discovery(or the intermediate ) server passes along client's details in which case yes, we will have 1 hop less
@@SangelRally I don't think so. How much microservice directly responds to the client. Also, how would the discovery service know if the request was completed or not?
Thanks for the video. I was wondering when your service communicates with ratings service we are not passing any token in the header. So how is the call secure or authorized call?
So how does Client knows that it needs to request for Service2 only.I mean if its already discovered , then it could have directly called the Service2 (Is it that the intermediate server just maintaining url mappings for the requests )
Hey thank you for your video, its really helpful. But i have one doubt. Let say eureka service is running one machine with IP address 10.2.162.12 and i start 3 different micro service and all 3 are running in different physical machine having IP address 10.2.162.13, 10.2.162.14 and 10.2.162.18 then how the service discovery will work. in all of example people show in same system, but its not the case in real deployment. in real deployment, to do horizontal scaling we will be running multiple small instance of service and they all can have different IPs. Please correct me if I misunderstood something.
Question : As when we make async call to a microservice using webclient , it does not wait for the response and moves on. What happens when we are performing some operation on the response after the call and we still havnt received the response from the other web services?
You do operation on response only after you receive it. If you are doing something ( during that async is happening) , that must not be related to or dependent on Response.
How the micro services are different from web services, in this tutorial you have created 3 different apps and each has web services. technically we are making rest call to communicate each other. i am so confused here why these are micro services
Question: we have spring mvc monolith application and we decided to break it and make microservices(spring boot) So now we have one monolith app and 2 microservices(from monolith app). We started using service discovery and registered microservices as services so that microservices could communicate without knowing ip of services. But we stuck how to call these services from monolith app as we trying to register monolith app as discovery service. Are we doing something wrong?
Hi Sir, I'm an intern SE from Sri lanka and I really like the way you educate us. Thank you.
Damn!! Indians are good in IT, wish we could have a better relationship with you guys, thanks for crystal clear information. Kudos from Pakistan.
wow. remineded me of my mathemetics professor lecture crystal clear touching on all loopholes/questions. gem of a person. :)
Between the client and the discovery server, there will be another service
which will redirect clients' calls to the discovery server.
This service usually called Gateway service.
Good explanation about spring boot
Hi,I think you should do these kinds of workshops more often with the students in person rather than creating youtube videos, this way we can get to know the questions and answers asked by the students in the video itself, and I must say some of the questions were very spot on, which clarified things furthur.
If you make a youtube video, we need to look at the comments to see if there is a questions/clarifications asked anywhere which is very difficult since there are many appreciation comments in between.
Thank you for sharing! Very helpful! Great explanation!
Great course and teaching. But I think you didn't explain the server side discovery pattern properly. The server side involves a load balancer in between and every request goes through it. And few other sites say it involves more hops.
concise and useful 🤓
He didn't take into consideration the response Hops which seem to be the issue in the comments. And Also the discover of the discovery server, which I believe will be picked up in future videos.
But other than that I felt the video was really good and on point in conveying exactly what was needed and not muddling one with unnecessary concepts.
Really Thank you for this series.
The response from the service in case of server-side discovery is given directly to client or it is relayed back to client using discovery server?
Can we use both API gateway and Service Discovery together? If yes how do we do that?
Good explanations thanks, actually understood how spring boot do things efficiently and magically
Can the discovery server handle load balancing as well?
For example, if the client wants to talk to service 2(there are more than 1 instance of it) can discovery server balance load among them.
Either URL is hardcoded or injected via some other configuration System ( For instance Chef).
@Java Brains, @Koushik - let's say I have one microservices which is build in another technology apart from springboot and I want to register that to springboot Eureka server ? Can we do that? Assume that microservices is in NODE ?
Thanks for sharing
I have one question ..
what if Discovery client goes down?
do we have any mechanism recover that. Bcz I find client side Discovery risky bcz everything than depends on one hope 🤨
I suppose you replicate discovery server, implement failover mechanisms and so on...
Your Service Discovery Serve should be highly available.
Maybe I missed it, but how did he start all 3 services at once?
Thanks so much for this tutorial.
What is the tool you are using for creating your presentation??
How are there no extra hops? There whould be no extra hops if there were no data to retun to the client.
On the server side case we got: Client -> Discovery -> Mircroservice -> Discovery ->Client
On the client case we got : Client -> Discovery -> Client-> Microservice -> Client
4 hops in each case except if on the 1st case when discovery(or the intermediate ) server passes along client's details in which case yes, we will have 1 hop less
In your first case "Client -> Discovery -> Mircroservice -> Discovery ->Client" , there is no need for the secondlast "Discovery".
@@SangelRally I don't think so. How much microservice directly responds to the client. Also, how would the discovery service know if the request was completed or not?
Thanks for the video. I was wondering when your service communicates with ratings service we are not passing any token in the header. So how is the call secure or authorized call?
It's not, in this case.
So how does Client knows that it needs to request for Service2 only.I mean if its already discovered , then it could have directly called the Service2 (Is it that the intermediate server just maintaining url mappings for the requests )
hi Kaushik on the server-side would it be fair to say that the discovery server is not a discovery server but something similar to a reverse proxy?
I guess you did not explain how to discover the discovery server. Can you explain, please?
Hey thank you for your video, its really helpful.
But i have one doubt.
Let say eureka service is running one machine with IP address 10.2.162.12
and i start 3 different micro service and all 3 are running in different physical machine having IP address
10.2.162.13, 10.2.162.14 and 10.2.162.18
then how the service discovery will work.
in all of example people show in same system, but its not the case in real deployment. in real deployment, to do horizontal scaling we will be running multiple small instance of service and they all can have different IPs.
Please correct me if I misunderstood something.
WHAT'S THE SOLUTION
I also dont know solution for this using service discovery of spring, but i feel Kubernets can help in this.
can you make video on microservices with jwt authentication
i like all your videos but one suggestion please use mic while recording a video.
what is difference between Discovery server and API Gateway,
Great Koushik ...
Question : As when we make async call to a microservice using webclient , it does not wait for the response and moves on. What happens when we are performing some operation on the response after the call and we still havnt received the response from the other web services?
You do operation on response only after you receive it. If you are doing something ( during that async is happening) , that must not be related to or dependent on Response.
In this tutorial, are we dealing with stateless or stateful?
As it's on Rest Principal , it's bound to be stateless
Is there a link to those slides?
How the micro services are different from web services, in this tutorial you have created 3 different apps and each has web services. technically we are making rest call to communicate each other. i am so confused here why these are micro services
You can say these as microservices if they form an application by connecting with other services. Otherwise these are like just a SpringBoot project.
Question: we have spring mvc monolith application and we decided to break it and make microservices(spring boot) So now we have one monolith app and 2 microservices(from monolith app).
We started using service discovery and registered microservices as services so that microservices could communicate without knowing ip of services.
But we stuck how to call these services from monolith app as we trying to register monolith app as discovery service. Are we doing something wrong?
How to do server side -service discovery in springboot?
Does anyone have any idea?
Not sure about Sprongboot but you can use etcd, consul nginx fo the same.
you silently skip this point 'how client will know where the discovery server is' ?
#for the algo
you sounds like Hritik Rosan
poor explanation on the server side service discovery.