- Newest
- Most votes
- Most comments
The issue you're experiencing with your Lightsail instance is likely related to CPU resource exhaustion. Let me address your questions:
-
Why does high CPU cause Status Check Failed? When your Lightsail instance consistently consumes more CPU than the plan provides as a baseline, it can lead to performance issues. Lightsail instances use a burstable performance model, which means they have a baseline CPU performance with the ability to burst above that baseline for short periods. Once these CPU credits are depleted, the instance's performance drops significantly, making it appear unresponsive and causing status checks to fail. The system becomes so overloaded that it can't properly respond to status check requests.
-
Why doesn't the instance recover automatically without reboot? When an instance enters this state of resource exhaustion, the operating system and services running on it may become deadlocked or severely degraded. The instance itself is still technically running, but it's in such a resource-constrained state that it can't recover on its own. A reboot forces all processes to restart and clears the deadlocked state, allowing the instance to function normally again until the cycle repeats.
-
Ways to prevent this:
-
Scale up your instance: If your application consistently requires more CPU resources than your current plan provides, consider upgrading to a larger Lightsail plan with higher baseline performance and more CPU credits.
-
Optimize your application: Identify and disable any unnecessary resource-consuming processes. For example, if you're running a WordPress site, certain plugins or custom themes might be consuming excessive CPU in the background.
-
Monitor your instance: Keep a close eye on your CPU usage and burst capacity using Lightsail's built-in monitoring tools to identify patterns and make informed decisions about resource allocation.
-
Set up an automated solution: While not ideal as a permanent solution, you could implement a system to automatically reboot your instance when it becomes unresponsive, using a combination of external monitoring and the AWS CLI or API.
The root cause is almost certainly that your instance is undersized for your workload. Addressing this by either optimizing your application or upgrading to a larger instance size would be the most effective long-term solution.
Sources
AWS lightsail need reboot every ten hours | AWS re:Post
Lightsail instance becomes unreachable from network | AWS re:Post
How can I reboot lightsail instance using cloudwatch or something? | AWS re:Post
Did you ever make any progress with this? I had (what seems like) a similar problem. I solved it by adjusting the Swap space on my instance. (An Ubuntu instance running several Wordpress sites.) It turned out that the underlying problem was actually running out of memory. I don't log in here very often but here's a post somebody else made about this https://blog.adrianrubico.com/post/Boosting-AWS-Lightsail-Performance-with-Swapon-to-Save-Your-Low-RAM-Instance
Thank you for your reply.
I previously read many articles saying this issue might be related to high CPU usage. I checked my server’s metrics and confirmed that before the problem occurred, the CPU had entered the burstable zone. However, the burst capacity was far from being used up. Even if it were fully used, I understand it should only cause CPU throttling, not make the server completely unreachable.
In my case, it seems similar to what you described — the server becomes unresponsive after entering the burstable zone, even though some CPU credits remain. This explanation makes more sense than what I read elsewhere. Still, I don’t quite understand why an increase in memory usage would necessarily lead to high CPU usage and then cause the server to stop responding.
After my server second incident, I took several preventive actions: I disabled an unnecessary service and added CloudWatch monitoring for memory usage, as I suspected memory might be related. Previously, the server stopped responding after running for 20 days and 13 days. After applying these measures, it has now been running for 12 days, and memory usage stays below 30%. I will keep monitoring to see if this mitigation is truly effective.
Based on my current understanding, I tend to believe this might be an internal issue with AWS Lightsail itself.
Anyway, thank you again for your reply. If my preventive measures do not work, I will try the methods you suggested.