Is workload tiering a good usecase for GKE MCS? MCS = Multi Cluster Services

[dheerajpanyam]

 I remember seeing workload tiering in this link few days back under the heading "Benefits of  GKE MCS"  but it got removed 2 days back. We have our apps split across 4 GKE clusters so I am assuming this means workload tiering and my question is does this fall under GKE MCS usecase?

My point is in addition to having GKE clusters spanning multiple regions, zones , workload tiering also is a good usecase for GKE MCS am i correct? Any sample setup?

[garisingh · Accepted solution]

Hi @dheerajpanyam  -

Splitting different types of services across clusters is definitely a valid use case for MCS.  The most common cases we see are shared services (instead of deploying them in every cluster), separation of stateful and stateless applications, separation of workloads with completely different lifecycles, team-based separation, etc.

MCS itself was definitely designed for any use case where workloads in one cluster need to talk to workloads in another and you prefer not to use load balancer services.

[garisingh · Accepted solution]

Hi @dheerajpanyam  -

Splitting different types of services across clusters is definitely a valid use case for MCS.  The most common cases we see are shared services (instead of deploying them in every cluster), separation of stateful and stateless applications, separation of workloads with completely different lifecycles, team-based separation, etc.

MCS itself was definitely designed for any use case where workloads in one cluster need to talk to workloads in another and you prefer not to use load balancer services.

[dheerajpanyam]

Thanks so much @garisingh . Any reference docs showing sample deployments using MCS? Let me also search GCP documentation.

[garisingh · Accepted solution]

The docs have an example or two, but you might find this blog useful:  https://medium.com/google-cloud/gke-multi-cluster-services-mcs-feels-like-magic-at-first-de39847554c...

[dheerajpanyam]

Thank you @garisingh 🙏