What is the execution flow when deploying a service mesh?

Meshery
1

I’ve seen the overall architecture diagrams, but am struggling to trace the execution flow when deploying a mesh. I think it might be something like this?

  1. deploy mesh (either UI or CLI)
  2. API POST request is sent from client to <meshery-server>/api/system/adapter/operation
  3. meshery-server forwards request to specified adapter
  4. adapter uses meshery-adapter-library to create new mesh object
  5. adapter uses meshkit to invoke Helm operation (INSTALL) referencing the new mesh object
2

As an initial data point, if this hasn’t come across your desk already - - Not specific to service mesh provisioning, but to MeshKit and Meshery Adapter Library, this post by @michael offers some insight - https://layer5.io/blog/meshery/introducing-meshkit-and-the-meshery-adapter-library

1 Like
3

The Meshery Architecture deck contains a number of sequence diagrams, but doesn’t yet have one covering service mesh provisioning (help wanted).

  1. through 4. - - spot on.
    5.1. adapter uses meshkit to invoke Helm operation (INSTALL) referencing the new mesh object.
    5.2. if Helm operation fails, adapter uses secondary deployment operation(s), which most of the time means using the specific service mesh’s CLI as the deployment mechanism.
1 Like
4

Thanks for the quick replies! This is exactly what I was looking for. The post you shared is extremely helpful as well.

1 Like