Website Outage Logic Tree

Website outage logic tree

This logic tree helps an on-call engineer narrow a website outage without jumping straight to a single suspected cause. It separates reachability from application health. DNS failure, load-balancer failure, and an unhealthy application require different evidence and different owners. Once the basic checks pass, traffic spikes and upstream dependencies become plausible causes. The leaves are deliberately phrased as checks rather than diagnoses, because incident work should follow current evidence. Adapt the labels to the actual architecture and link each leaf to a runbook. A logic tree is useful when several causes can produce the same symptom. It gives readers a concise reference they can reuse when discussing the result with colleagues.

種類 決定木規格 Howard-Raiffa / CART-sklearn / taxonomyエンジン schematex-decisiontree更新日 2026/9/24

依頼内容

“A logic tree for a website outage: split into reachability and application checks, then include DNS, load balancer, health checks, traffic, deployments, logs, and database connectivity.”

次はこちらAdd a branch for expired TLS certificatesAdd a cache and CDN check

この図に含まれるもの

図の背後にある意思決定を読み解く。

01Reachability

Establish whether users can resolve and reach the service.

02Health

Separate infrastructure reachability from application health.

03Evidence

Use monitoring and logs at each leaf.

Use as a first-pass on-call troubleshooting guide for a web service.

決定木のテンプレートをすべて見る →

よくある質問

Why check DNS first?

A DNS problem can prevent users from reaching an otherwise healthy service.

Are branches always exclusive?

Not always; use evidence to identify which hypotheses need testing.

What belongs at a leaf?

A test, observation, or next diagnostic action.