Product Update: Monitoring EC2 and ALB with Anomaly Detection

We are glad to announce a product update: marbot uses anomaly detection to when monitoring an EC2 instance or Application Load Balancer (ALB).

Please note, that this product update does only affect customers, who are using the Monitoring Assistant, that configures AWS monitoring automatically.

Who is already using the Monitoring Assitant?

• All teams that installed marbot for Slack after 2022-12-15.
• All teams that installed marbot for Microsoft Teams after 2023-01-17.
• All teams that opted-in by sending an email to support.

Monitoring ALB with anomaly detection

We are removing the CloudWatch alarm marbot-target-5xx-count-too-high-*, that was monitoring server errors by target group to reduce costs caused by CloudWatch alarms. Nevertheless, marbot still monitors server errors on the ALB level with the CloudWatch alarm marbot-alb-5xx-count-too-high-*.

Formerly, the marbot-alb-5xx-count-too-high-* alarm fired when detecting any alerts via the HTTPCode_ELB_5XX_Count CloudWatch metric. From now on, marbot configures CloudWatch anomaly detection to monitor the HTTPCode_ELB_5XX_Count metric. This will minimize the number of false alarms, but will ensure that an alert is sent in case of unusual error patterns.

Monitoring EC2 with anomaly detection

By default, marbot uses anomaly detection to monitor the following metrics of an EC2 instance.

• CPU Utilization
• Network Utilization

Using anomaly detection instead of hard-coded thresholds reduces false alarms and makes sure you are getting notified if an EC2 instance runs under unusually high load.

Take your AWS monitoring to a new level! Chatbot for AWS Monitoring: Configure monitoring, escalate alerts, solve incidents.

Andreas Wittig

Consultant focusing on Amazon Web Services (AWS). Entrepreneur building marbot.io. Author of Amazon Web Services in Action, Rapid Docker on AWS, and cloudonaut.io.

You can contact me via Email, Twitter, and LinkedIn.

Published on 03 Mar 2023