How does everyone feel about service graphs in a brownfield network centric cutover? maybe I'm an old school engineer but I was taught that if you have to do a Pbr then your design is wrong. they are a pain to troubleshoot because you always forget they are there. With service graphs it seems you could quickly have a couple hundred of these. Once your apps goes productions they are then dependent on the PBR to complete the flow which makes migration to something that follows the evpn vxlan standard difficult. We are in the process of planning our deployment and I'm currently leaning to do either bridge domain or vrf stitching. How did you guys do it?
**Disclaimer**, I haven't deployed one yet I've only read the documentation
