Out of process adapter deployment

Policies and Telemetry
1

Hello everyone and thank you for your help.

I was able to complete the first functional draft of an out of process Mixer adapter that sends telemetry to my backend. The guide, however, doesn’t go into any detail regarding deployment of such adapter beyond local development. What are the recommended strategies for this? Do I tell my users to modify Istio’s deployment YAMLs to run my adapter on the same POD as Istio’s Mixer as a sidecar? Do I set it up as a separate Replica Set which scales independently and configure Mixer to talk to it over the service mesh? What’s the general recommendation here?

Thank you.

2

I am looking for the same recommendation. Any pointers will be appreciated.

3

Here’s a tutorial I had written describing the process.

https://venilnoronha.io/set-sail-a-production-ready-istio-adapter

Thanks,

Venil

4

Thank you. I kind of did the same thing intuitively, except my adapter is under .NET Core. Also, I believe the “critical” annotation you use is now deprecated, something to consider.