In response to a major outage on Monday, AWS has added an automated incident-generating capability within its CloudWatch service for reporting use cases.
CloudWatch itself is a monitoring and observability service that is targeted at helping enterprises get insights into the operational health of their AWS cloud services as well as respond to any changes for optimization.
The new capability, embedded within CloudWatch’s generative AI assistant CloudWatch investigations, is designed to help enterprises create a comprehensive post-incident analysis report quickly.
“The new capability. automatically gathers and correlates your telemetry data, as well as your input and any actions taken during an investigation, and produces a streamlined incident report,” AWS wrote in a blog post.
These reports will include executive summaries, timeline of events, impact assessments, and actionable recommendations, helping enterprises to identify patterns, implement preventive measures, and continuously improve their operational posture, AWS added.
Forrester principal analyst Charlie Dai said the new capability is a way for AWS to regain trust of its customers, especially after the outage, which was later traced to a malfunctioning DynamoDB endpoint.
These reports can be effective for enterprises in improving their resilience, Dai said. However, he did point out that AWS could better help customers minimize downtime and business risk by promoting multi-region architectures, active-active failover, and redundant DNS strategies.
Further, he said that while reports will help in accelerating post-mortem analysis, it is far from enough, and only continuous product improvement, along with practice optimization, can help minimize systemic risks.
Generating an incident report
In order to take advantage of the new capability, enterprise users need to ask questions off the CloudWatch investigation assistant about a particular service’s performance issues or the reason behind its downtime.
Once a user requests such information, the AI-powered assistant scans the system to find telemetry that might be relevant to the situation, and generates hypotheses based on what it finds.
Once the hypotheses are accepted by the user, the assistant can be asked to generate an incident report, the company wrote in its documentation.
Currently, the incident report generation feature is available in US East (N. Virginia), US East (Ohio), US West (Oregon), Asia Pacific (Hong Kong), Asia Pacific (Mumbai), Asia Pacific (Singapore), Asia Pacific (Sydney), Asia Pacific (Tokyo), Europe (Frankfurt), Europe (Ireland), Europe (Spain), and Europe (Stockholm) regions.
AWS’ outage has also spiked interest from other observability vendors, with Datadog launching a free website for enterprises to monitor the status of services from multiple cloud service providers.
However, Datadog’s website isn’t the only one of its kind; similar sites, especially aggregators of status pages and user reports-based trackers, such as Updownradar.com, IsTheServiceDown.com, and Downdetector, already provide information on outages. Nearly all cloud service providers, such as Google, Microsoft, and Alibaba, provide a service-related status information page or service: Azure Service Health provides personalized alerts, root-cause reports, and guidance during incidents; Google Cloud offers Service Health dashboards and custom alerts for affected resources; and Alibaba Cloud has Incident Response Service for emergency handling and post-incident planning.
Leave a Reply