Integrations/spark
Already run spark? We read it automatically.
clan.me soft-depends on spark. When it's installed, your dashboard shows spark's trusted TPS, MSPT percentiles, and system CPU, with a “via spark” tag so you always know the source. No config, no restart needed.
What spark adds
Four things you only get with spark installed
Trusted TPS
clan.me reads directly from spark's TPS API, the same source as /spark tps. No more discrepancies between what spark reports and what your dashboard shows.
MSPT percentiles
p95 and p99 tick times alongside the mean. Catches spikes that an average smooths over, shown as sub-values in the Snapshot panel.
System CPU %
System CPU utilisation from spark's cpuSystem API, a metric clan.me can't collect without spark. Charted over time in Server Health when spark is present.
MSPT on Spigot
MSPT normally requires Paper's API, which Spigot doesn't expose. When spark is installed on Spigot, clan.me reads MSPT from spark, a genuine capability gain.
In your dashboard
The continuous record a single profile can't give you
A recurring nightly dip, only visible across the 7-day chart.
19.6
19.8 / 19.5 (1m/15m)
TPS (5m)
32.4ms
p95 41.2ms
MSPT
18.3%
CPU
The Snapshot panel with spark-sourced TPS, MSPT p95, and CPU.
Auto-detection
Zero config. Just works.
Server starts
The clan.me plugin checks for spark at enable time, via the plugin manager or the bundled spark inside Paper 1.21+.
spark detected
If found, SparkBridge initialises once and the console logs "Metrics source: spark". Nothing to configure.
Every heartbeat
TPS, MSPT, and CPU are read from spark's live API. The "via spark" tag appears in your Snapshot panel.
Nothing changes if spark isn't installed. The plugin falls back to native measurement silently. Sessions, retention, entities, GC, alerts, and every other feature work exactly as before.
How they fit together
clan.me is the alarm. spark is the microscope.
clan.me
- ·Continuous record, every 60 seconds
- ·30 days of raw pings, 18 months of hourly aggregates
- ·Alerts at 3am when you're not watching
- ·Tells you when to run spark
/spark profiler
- ·30-second point-in-time profile
- ·Method-level call stack attribution
- ·You run it manually, right now, on demand
- ·Tells you what is causing the problem
You run spark when you already know something's wrong. clan.me is what tells you something's wrong, before your players notice. They are complementary tools aimed at opposite ends of the same problem.
Setup
Already have spark? You're done.
If you're running Paper 1.21 or newer, spark is bundled internally and clan.me detects it with no extra steps. If you're on an older version or Spigot, install spark as a standalone plugin first.
Paper 1.21+
spark is bundled. Install the clan.me plugin, add your token to config.yml, restart. No separate spark plugin needed.
Older Paper or Spigot
Install spark from spark.lucko.me, then install the clan.me plugin. Both detect each other automatically on next start.
Common questions
Do I need spark for clan.me to work?
No. spark is completely optional. Roughly 70% of clan.me servers run without it and still get full TPS, memory, GC, entity counts, sessions, retention, and all eight alert types. spark just adds CPU monitoring, MSPT percentiles, and trusted TPS on top.
Paper 1.21+ bundles spark internally. Do I need a separate plugin?
No. clan.me checks for the bundled spark too. If you're on Paper 1.21 or newer and haven't disabled spark, clan.me picks it up automatically with no separate install.
Will TPS numbers match what /spark tps shows?
Yes. When spark is detected, clan.me reads from the same spark API that /spark tps uses. The numbers are identical because they come from the same source.
Which spark versions are compatible?
Any version of spark compatible with your server. clan.me uses the stable spark developer API (SparkProvider), which has been consistent across spark releases.
Other integrations: clan.me also auto-detects BlueMap, Dynmap, and Squaremap for live map thumbnails on your server page. See all integrations →
Start monitoring your server.
Install the plugin in 60 seconds. If you already run spark, the integration is automatic.