Troubleshooting Performance IssuesThis page outlines troubleshooting steps you can take if you're experiencing performance issues with Cryptowatch.
If notice an extremely low fps (frames per second) on the Cryptowatch chart, your browser may be misconfigured.
This is probably due to hardware acceleration being disabled for HTML5 canvas. Enable it and restart your browser.
- In a new tab, open
- Make sure
- Restart Firefox
You may also have to set
true. Read more.
Incidentally, Firefox 57 introduced an update which can dramatically speed up page rendering. Upgrading to Firefox 57 might be a good idea.
Lastly, we notice that in general Google Chrome has an easier time hitting 60fps when scrolling/moving cursor. We recommend Chrome over Firefox.
Open Chrome Settings and find Advanced > System
Make sure "Use hardware acceleration when available" is enabled
In a new tab, open
Make sure "Accelerated 2D canvas" is enabled
High CPU Usage
If the chart and interface are not lagging but Cryptowatch is consistently using a lot of CPU (50% or more), you can try a few things:
- In the settings panel, throttle the order book refresh rate
- In the settings panel, make sure "animate feeds" is disabled
- Draw fewer indicators or make candles wider
- Run fewer Cryptowatch tabs
- Buy a better computer
If none of the suggestions above help, you can reach out to us with some diagnostics to get some help. It's possible you've stumbled across a bug or inefficiency in our code that is causing the issue.
Record a performance profile while you are experiencing the high CPU usage
Found under the "Performance" tab in Chrome or Firefox developer tools
Save the profile file and email it to email@example.com, along with the following details:
Which browser(s) you're having problems in
Which URL(s) which exhibit the problem
How consistently the problem happens (occassionally? every time?)
How long the Cryptowatch session is open before the problem starts (after a few days? immediately?)
Our developers can use this profile file to better understand where the bottleneck is and fix it in our codebase.