This guide will show you how to install Spark and teach you about some of its main features.
Spark is a performance profiling plugin/mod which provides you with a variety of tools such as diagnosing performance issues, diagnosing memory issues and keeping track of overall server health. While Spark is simple to install it has a lot of options and commands that can feel a bit overwhelming to new users.
- Download the Spark plugin for your Forge/Fabric server or Spigot/Bukkit server.
- Once downloaded, connect to your server via FTP. You can do this either using the web FTP, or by using an FTP client such as FileZilla.
- Upload Spark into the
modsfolder, depending on your server type.
- Once the upload is complete, start your server.
Main Spark Commands
/before any of the commands.
||Starts the profiler in default mode.|
||Checks current status of profiler.|
||Stops the Profiler and sends a link to view results. Copy and paste the link into your browser to view.|
||Prints information about the servers Ticks Per Second rate and CPU Usage.|
||Generates a health report for the server including TPS, CPU, Memory and Disk Usage.|
||Prints information about recent activity performed by Spark such as when you last ran the profiler and the link for the last profiler ran.|
||Performs a short scan of your server ticks to find your average tick rate. From then on, any ticks with >100% increase in comparison to your average will be reported.|
Listed below are some examples of what the outputs of commands look like both in-game and in your Multicraft console.
How to Analyze the Profiler
The profiler is very similar to a timings report. Once you obtain your link after stopping the profiler, you can copy and paste it into your browser. This link can be sent to Support to help our staff understand the cause of lag however, it can be insightful to you as well.
- To view more information, keep clicking the options that appear to show an in-depth breakdown of what they are. You’ll want to pay attention to the percentages beside each task.
- From there, locate the processes that take up a lot of the servers resources. In this case, these are usually things that are running frequently. In the example below, we can see that "updateTimeLightandEntities" is taking a lot of time to do (the percentage is fairly low), expand it to further investigate.
- You’ll eventually narrow it down to one or two things that are causing high usage (lag). In this case, "guardEntityTick" and "ServerChunkProvider.tick". We can tell that these are higher than normal, as Spark has a helpful feature that allows you to see what a "normal" amount of time for these processes take - simply hover over the option to show the normal time spent on the task.
- It’s useful to note you can also bookmark tasks by right clicking on them and clicking "Toggle Bookmark". This will update your profiler link, so when you show it to our staff, they will know exactly what you're talking about.
To find out more on Spark, you can find a helpful guide on More Commands and their Usage.If you require any further assistance, please contact our support at: https://shockbyte.com/billing/submitticket.php