How to avoid "Positive Feedback Cycle Overload Problem"?

Question

Sometimes while designing reliable systems, we try to make the system more reliable by adding retries in event of failure (with feedback mechanisms). And it results to potential for an overload because we may be adding more load to an already overloaded system. How intelligent retries can be done considering overload conditions?

Answer 1

Depending on the infrastructure in your project you could create an autoscaling policy and depending on how you define it, your system will be affected. Please take a look at the following documentation that may help you to get a better understanding on how you can implement a good autoscaling policy.

Autoscaling groups of instances

Load balancing and scaling

Reliability