Resto Druid parses live and die on HoT uptime more than any other single stat, but "keep your HoTs up" isn't specific enough to actually act on. Here's what the uptime, mana, and overhealing numbers in your log are actually telling you, and how Healper weighs each one when it scores a Resto Druid run in Mythic+.
What top Restoration Druids actually did
Measured from 400 logs of top-performing Restoration Druids at keystone level 11 in Midnight Season 2 (Curse of Ula'tek), across 8 dungeons. These are measurements of what that cohort did, not targets anyone set - and the middle-half column matters more than the median, because it shows how much honest variation there is between good runs.
| What | Median | Middle half | Logs |
|---|---|---|---|
| Rejuvenation uptime | 96.4% | 94.4-97.8% | 400 |
| Lifebloom uptime | 88.1% | 81.5-93.0% | 400 |
| Wild Growth uptime | 43.6% | 37.5-49.4% | 400 |
| Average mana through the run | 71.8% | 61.2-81.6% | 400 |
| Overhealing | 59.2% | 55.2-62.6% | 400 |
11 of those 400 runs had the healer bottom out on mana at some point (2.8%). Running dry is not rare at this level, and the spread in the mana column is why: the middle half of this cohort is a wide band, not a target.
The same corpus broken down per dungeon, with the cooldowns this spec actually pressed, is on the Restoration Druid benchmarks page. Mythic+ only: the raid brackets carry no per-log data in any season, so nothing here describes raid.
The uptimes that carry the most weight
Not all HoTs are worth the same: Lifebloom carries the most of the uptime score, then Rejuvenation, then Wild Growth, which is a burst tool rather than something to maintain. The exact weights live on how it works rather than here, so there is one copy of them and it cannot drift.
Uptime is calculated as the percentage of active combat time - gaps of more than 3 seconds between events count as downtime and are excluded - during which the HoT is active on at least one target in your group. It's not per-target uptime, and it accounts for pandemic refreshing: reapplying Lifebloom just before it expires preserves part of the remaining duration instead of resetting the timer, so a well-timed refresh doesn't cost you anything.
The practical takeaway: losing Lifebloom uptime is the most expensive mistake you can make, simply because it's worth the most points and it's the HoT most likely to fully drop during a hectic pull. If you're triaging where to focus, start there before worrying about squeezing out more Wild Growth casts.
Mana: don't starve, don't overcast
Mana is graded on two things: your average mana % across the fight, and whether you went fully OOM at any point. Going OOM costs half of whatever your mana score would otherwise have been - it's treated as a distinct failure, not just "low mana," because running out mid-key usually means a bigger pacing problem than a single low reading does.
Overhealing isn't free
The last block of the score is overhealing %, compared against benchmark logs at your bracket rather than a fixed target. Healing that exceeds a target's missing health is "wasted" from a scoring perspective even though the cast itself wasn't wrong - the signal worth paying attention to is a persistently high overheal rate relative to top performers at your key level, which usually points to casting reactively (topping off full-health targets) instead of reading incoming damage.
Boss mechanics: pre-cast, don't react
None of the above captures timing relative to the fight itself. Healper separately checks whether you used the right tool before a dungeon's known dangerous mechanics - for example, pre-casting Wild Growth a few seconds ahead of a scripted burst-damage mechanic, the way top performers in that dungeon do - rather than just reacting to the damage once it lands. This is compared against what top performers actually did on the same pull, not a generic "always pre-HoT" rule.
All of this - the exact percentile thresholds, the pandemic-refresh math, the scoring formula - is documented in full in how we calculate your score.