Welcome to the Community,
for a single ESXi host (i.e. without Distributed Virtual Switches) you'd usually use the vSwitches/Portgoups with their default Teaming&Failover settings. This will allow you to connect the uplinks to different switches without complex configurations, and the uplinks will be assigned to the VMs at power on in a Round-Robin manner for load distribution.
If multiple VM portgroups with different VLANs are required on the ESXi host, then go with untagged (802.1Q) ports on the physical switches, and configure the VLAN-IDs on the virtual portgroups.
This choice is highly dependent to your network requirement and VLAN design, I mean if you consider all servers inside your environment in a single network portion regardless of they are physical system or virtual machine, you can ignore to setup VLAN ID on port groups, and configure the physical uplinks as the access port in the physical switch. But if you need to isolate L2 traffic inside the virtual networking based on design, for example if you need to separate frames that are generated via infra VMs from database VMs and so on, so you need to set VLAN ID for each of port groups and then you should put the corresponding pNICs in trunk mode in the Phy-Switch.
About the using of link aggregation just one point you should be aware of: If you need to configure a LACP port inside the physical switch contains VMNICs of a virtual switch, you should set the load-balancing mode of that vSwitch (inside the policy exceptions in the NIC teaming tab) in the IP-Hash options.Please mark my comment as the Correct Answer if this solution resolved your problem
Moderator: Moved to vSphere vNetwork