Incident Severity Classification

Report an Issue

Help us improve by telling us what went wrong.

Embed this Tool

Customize and copy the embed code for your website.

Classify Incidents Consistently Across Your Organization

When a production system goes down at 2 AM, the first question everyone asks is: how bad is it? The answer determines who gets paged, how quickly teams mobilize, and what communication goes out to customers. But without a clear classification framework, severity assessments become subjective, varying wildly between on-call engineers, teams, and shifts. The Incident Severity Classification Tool provides a structured, repeatable framework for assessing incident severity so that every engineer in your organization makes consistent, defensible classification decisions.

Inconsistent severity classification causes real damage. Over-classification triggers alert fatigue and erodes trust in your escalation process. Under-classification delays response to genuinely critical issues and leaves customers without timely communication. This tool eliminates both failure modes by grounding classification in objective criteria.

The Classification Framework

The tool uses a multi-factor assessment model that evaluates incidents across several dimensions. User impact scope measures what percentage of users are affected: all users, a significant segment, a small group, or internal users only. Functionality impact assesses whether the affected capability is a core revenue function, an important but non-critical feature, or a convenience feature.

Data integrity checks whether data has been lost or corrupted, which elevates severity regardless of user-facing impact. Security implications determine whether the incident involves unauthorized access, data exposure, or compliance violations. Duration and persistence considers whether the issue is intermittent, sustained, or escalating.

Each factor is scored, and the tool produces a recommended severity level from SEV-1 (critical, all hands on deck) through SEV-4 (minor, fix during business hours). The scoring is transparent, so you can see exactly which factors drove the classification and adjust if circumstances change.

Using the Tool During an Incident

Speed matters during incidents. The tool is designed to be used in under two minutes. The on-call engineer answers a series of quick, unambiguous questions about what's happening. There's no room for philosophical debate, just factual assessment of observable impact.

Once classified, the tool outputs the severity level along with the expected response: who to notify, what the expected response time is, whether a status page update is needed, and whether customer communication should go out. These response expectations are customizable to match your organization's incident management process.

If the incident evolves, as they often do, run the classification again with updated information. A SEV-3 that started as a minor latency increase might become a SEV-1 when the root cause turns out to be data corruption spreading through your database.

Who Should Adopt This Tool?

On-call engineers are the primary users, making classification decisions under pressure with incomplete information. Having a structured framework reduces the cognitive burden and gives them confidence that they're making the right call.

Incident commanders coordinating response efforts use the classification to set expectations with stakeholders. A SEV-2 classification communicates urgency without the panic of a SEV-1, helping everyone understand the situation quickly.

Engineering leadership reviewing incident retrospectives can verify whether initial classifications were accurate, identifying patterns where incidents were consistently over or under-classified.

SRE teams establishing incident management programs can use the tool's framework as a starting point for their severity matrix, customizing the criteria and response expectations to match their organization's specific needs.

Practical Applications

A SaaS company adopts the tool and discovers that their team had been classifying payment processing failures as SEV-2 when the framework clearly indicates SEV-1 due to direct revenue impact. Upgrading the classification triggers faster response and reduces mean time to resolution for these critical incidents.

A platform team uses the tool to create a shared severity language across fifteen microservice teams. What one team called "critical" and another called "high" now maps to the same SEV level with the same response expectations.

Best Practices for Incident Classification

Classify based on impact, not cause. The root cause might be trivial, but if it's affecting all users, it's still a high-severity incident.

When in doubt, classify higher and downgrade later. The cost of under-classifying a real emergency far exceeds the cost of briefly over-mobilizing for a minor issue.

The Incident Severity Classification Tool runs in your browser with no data transmission. Your incident details remain confidential, and the tool is available around the clock with no account required.