AWS Elastic Beanstalk Support
Elastic Beanstalk environment health Severe or Degraded: how we fix it.
Beanstalk enhanced health turns red when requests fail, instances stop responding or a deployment goes wrong. The health Causes list and the instance logs usually point to the problem. We trace it from the environment to the instance and fix the application or configuration behind it.
When You Need This
The situations where this work pays for itself.
Signs it is time
- The environment health is Severe, Degraded or Warning
- A deployment ends with Command failed on instance
- The load balancer returns 5xx and instances are marked Unhealthy
- Health flapped after a configuration change or platform update
- Instances are stuck or repeatedly replaced by Auto Scaling
What you get
- Cause of the health state, with log evidence
- Corrected configuration or application fix
- Safer deployment policy settings
- CloudWatch alarms for environment health
How we do it
A clear, safe process you can follow along with.
1. Read the health Causes and events
In the console Health page, or eb health --refresh, we read the Causes column, such as 5xx responses, instance not responding or command failed, and the Events tab for the last deployment and configuration activity.
2. Pull full logs
We request the full log bundle (eb logs or the console) and read /var/log/eb-engine.log and eb-hooks.log for deployment errors, web.stdout.log and the app log for application errors, and the nginx or Apache error and access logs for proxy problems.
3. Check the health check path and load balancer
We confirm the application health check URL returns success for the load balancer, that the application listens on the port the proxy expects, and look at the ALB target group health and 5xx metrics. Health check failures from slow startup or auth redirects are common.
4. Look at instance resources
We check CPU, memory, disk and CloudWatch metrics. Full disks from logs, memory exhaustion in small instances and CPU credit exhaustion on T instances all push the environment to Degraded.
5. Inspect recent changes
We compare the last working application version and environment configuration, including .ebextensions and platform hooks, environment variables and the platform version, with the current one. A failing container_commands step or a missing env var is typical.
6. Fix and redeploy safely
We correct the cause, and use an immutable or rolling with additional batch deployment policy, so a bad version does not take capacity away. If needed we redeploy the last known good version first to restore service.
7. Add visibility
We enable enhanced health log streaming to CloudWatch, set alarms on environment health and 5xx, and document the runbook.
Mistakes we help you avoid
Common ways this goes wrong, and what we do instead.
Rebuilding the environment first
It hides evidence and may fail the same way. Logs come first.
All-at-once deployments in production
A bad version takes the whole environment down.
Editing instances directly
Changes are lost when Auto Scaling replaces the instance. Use configuration in the source bundle.
Ignoring disk space
Logs filling a small root volume cause odd failures.
Frequently Asked Questions
What does Severe mean in Elastic Beanstalk?
It is the most serious health color. It typically means most requests are failing or instances are not responding, and the Causes list explains the trigger.
How do I get logs from a Beanstalk environment?
Request the full or the last 100 lines of logs from the console or eb logs, or stream them to CloudWatch Logs. Key files are eb-engine.log, web.stdout.log and the proxy logs.
Why did my deployment fail with Command failed on instance?
A step in your deployment, such as a container_command, hook or package install, returned an error. eb-engine.log shows the command and its output.
Want Help With Elastic Beanstalk environment health Severe or Degraded: how we fix it?
Schedule an appointment. Tell us about your environment and we will outline the safest, highest-value approach.
Schedule Appointment › No commitment required