Combat is part of the game.
If a spell is better than some out of combat, and worse than them in combat, that is part of how 5e balances options.
Out of combat, if this was a 6th level spell as you think it should be, it would have to be beefed up noticably to not be noticably weaker than nearly every other spell of its level.
As it is, it’s in line with other 2nd level spells.
Is it? Other second level healing spells are... well, prayer of healing, which can't be used in combat at all and is significantly worse out of combat, and upcast 1st level cure wounds and word of healing. So, then an upcast cure wounds is going to hit for 2d8, which is better than 1d6, but then stop. So, for a round after round long fight, Healing Spirit is better, for short, quick fights, upcast Cure Wounds is better. Word of Healing deals with an entirely different angle -- it doesn't really heal damage so much as get people up from 0, and the odds of it being upcast are pretty slim outside of having no 1st level slots available. Even then, it's not better than Healing Spirit, which can do the same but then continue to heal allies, outside of the action economy. So, then for the other 2nd level healing spells in combat, Healing Spirit is clearly better than Prayer, and conditionally better or worse than the other two healing spells.
Out of combat, it's absolutely no contest. Healing Spirit is far and away a better choice than any of the other available 2nd level healing spells. Not even an contest.
So, then Healing Spirit is competitive in combat, and clearly superior out of combat. The question remains, does it unbalance the game at all because of this?
You keep asking for a point at which it would actually unbalance a real game in play. Well, I offer my game as an example. Recently, the party finished an encounter and knew it couldn't take the time to rest and that additional encounters were likely. This pushed the party cleric to using his 2nd level slots (the party is 3rd level) to upcast some cure spells to deal with the large amount of damage 2 of the characters had taken. This made the choice to continue a risk, as the cleric no longer had those slots available to contribute. Had the party druid been able to use 1 2nd level slot to heal everyone (something that didn't happen, as 3 of the 5 characters carried damage into the next fight), that choice wouldn't have been made, and the party would have been able to continue with 1/2 the spell resources spent and all of their hitpoints returned. Could I modify the encounters to account for this? Sure, but if I have to modify the way I present encounters and their pacing based on whether or not there's a caster of Healing Spirit in the party, then I can easily say that Healing Spirit distorts the game all on it's own. Being able to counter it doesn't mean that it's not a problem. I'm the DM, I can counter anything I want. The question is: should I have to make as big a change as healing spirit requires and only if it's present?
Further, as I'm running a sandbox, altering the game in that manner isn't something I'm willing to do merely because Healing Spirit is present. So, it will very much alter how my party engages what was built - prior to XGtE - as an exploration and survival based game by making survival far less of a challenge while exploring. Unless I change it up because HS is present.
HS is a problem for my game. I've solved it by saying it only works once per round- so only 1 healing per round no matter how many people wander through it. The caster controls when it heals, so it's up to them who gets the goodies. This makes it still better than Prayer of Healing, but not by much, and still lets it be useful in combat, but, again, not as abuseable by tactically minded players (like mine). Further, it upscales by 1d6 every 2 levels, not every level. This helps me as well, as my Druid player can contribute to healing in a useful way, but not totally outshine my Life cleric player.