- Newest
- Most votes
- Most comments
Worth to take a look at this:
- Check Health Check configuration
- Review ALB Target Group Health Checks
- Verify Security Group & Firewall rules
Hey,
Hope you're keeping well.
Global Accelerator marks endpoints unhealthy based on its own health checks, which are separate from your ALB target group checks. In the console, open your endpoint group and review the configured health check protocol, port, and path to ensure they match what the ALB can serve from outside its VPC. For an internal ALB, you’ll need to make sure the Global Accelerator health check source IP ranges are allowed in the ALB’s security group, otherwise the checks will fail. You can also use VPC Flow Logs or ALB access logs to confirm whether the health check traffic is reaching the ALB and returning a 200 response.
Thanks and regards,
Taz
answered 10 months ago
Another useless non-answer. You clearly didn't read the question and previous answers and comments.
The issue is that the ALB Health checks don't (or didn't at the time) state exactly WHAT specifically, the ALB Health checks complained about. Also AWS documentation at the time stated that if Global Accelerator pointed to an AWS ALB, then the Global Accelerator was not supposed to perform Health checks different from those done by the ALB .
Furthermore, your answer completely fails to state (directly or by pointing to a specific documentation subpage) what that supposed "Global Accelerator health check source IP ranges" are. Also your sentence about ALB and "outside its VPC" seems to have the direction reversed or a noun missing. In particular, the AWS Global Accelerator documentation at the time of the question talked about pointing the Global Accelerator to an INTERNAL ALB, while what your describe is an undocumented requirement to point only to a PUBLIC ALB.
Anyways as a matter of workarounds, I found that I had to make sure every unused ALB backend was running, even when the ALB itself was explicitly set to use the stopped instance 0% of the time.
Relevant content
asked a year ago
asked 3 years ago
asked 2 years ago
This is such a vague and useless answer. Randomly checking every related item is a hopeless strategy for troubleshooting anything. Desire is to make the AWS code TELL me what IT objects to. Also, it is well documented that the "Health Check configuration" of a GlobalAccelerator is ignored when pointing to an AWS ALB . The ALB itself reports no health problems (at least in the AWS Web console), which makes it even more important to find out why the GlobalAccelerator sees things differently.