AP_BattMonitor: extend AP_BATT_MONITOR_MAX_INSTANCES to 16

[magicrub]

AP_BattMonitor: extend AP_BATT_MONITOR_MAX_INSTANCES limit to 16 (default is still 9)

Doesn't hurt to bump the max up to 16, especially for vehicles with lots of power monitoring and AP_Periphs with sensors galore!

Note: if we ever consider pushing above 16 The limitation is param1 being treated as uint16 in handle_command_battery_reset()

MAV_RESULT GCS_MAVLINK::handle_command_battery_reset(const mavlink_command_long_t &packet) {
    const uint16_t battery_mask = packet.param1;
    const float percentage = packet.param2;
    if (AP::battery().reset_remaining_mask(battery_mask, percentage)) {

Since it's a command_long, the float could give us 24 bitmask bits as-is. The MAVLink docs don't mention any limit but since packet.param1 is a float we could change it to 32bit and do:

const uint32_t battery_mask = is_equal(packet.param1, -1.0) ? 0xFFFFFFFF : packet.param1;

which would allow a GCS to clear them all (including SITL autotest) without having to know how many there are and thus support up to 32 gracefully-ish.

In fact, I've also made PR #24661 on top of this branch that I made for the LOLs.

[magicrub]

Regarding PR #24645 where I bumped up RPM's instance count from 2 to 9, it was argued that it shouldn't be that high "because no one is asking for it" and that was a fair argument because I was only in need of 4 (at least for now). For this battery PR, I'm asking for 16 because I have a need for 12 right now and will likely need 13-14 pretty soon and I anticipate 16 eventually.