With the move to serverless architecture and auto-scaling cloud services, some of my colleagues argue that traditional Problem Management is obsolete. I disagree, but I'm looking for solid arguments or use cases on how it evolves in a modern, cloud-native ITSM setup.
3 answers
Problem Management is more critical than ever in the cloud, but the focus shifts from "fixing a server" to "fixing the code or configuration." In a serverless world, incidents are often transient, but the underlying 'Problem' could be a poorly written Lambda function or an API rate limit issue that will recur. You need to perform Trend Analysis on cloud spend and performance spikes. ITIL 4 suggests that Problem Management should work closely with Site Reliability Engineering (SRE). Instead of just Root Cause Analysis, we now focus on 'Error Budgets' and 'Blame-Free Post Mortems' to improve system resilience.
Do you find that your cloud providers give you enough granular data to actually perform a deep-dive Problem Analysis, or are you often stuck waiting for their internal teams to acknowledge an underlying platform issue that is out of your direct control?
It definitely evolves into "Proactive Problem Management." It’s less about reacting to outages and more about identifying architectural weaknesses before they cause a failure.
Spot on, Susan. Moving from reactive to proactive is the hallmark of a mature ITSM organization, especially when dealing with complex cloud microservices.
Brian, that's the biggest hurdle. We use third-party observability tools to get the telemetry the cloud provider doesn't show. This allows us to prove where the fault lies. Even if it is the provider's fault, Problem Management helps us design "Circuit Breaker" patterns in our own code to mitigate the impact next time, which is a proactive step.