Client profile
ShieldEdge, a US cloud security platform with 160 staff.
Challenge and scope
A daily queue of 3,800 alerts includes many false-positive escalations. Reviewers have to identify critical cases quickly across four product modules.
A single event may trigger several alerts across monitoring tools. Responders need to know which signal is current, how much of the service is affected, and whether the alert needs immediate attention. Overlapping events can make a routine notification look urgent or obscure a genuine critical incident.
Delivery workflow
Apply approved severity and routing rules to the incoming queue around the clock. Manually triage alerts assigned to the desk, document first-line findings, and hand critical incidents to the client’s responders.
Group related events when the client’s rules allow it and record the alert ID, detection time, impacted service, and initial severity. Use approved dashboards to validate the signal and link the supporting traces. Notify the designated responder with the evidence and capture acknowledgment in the incident record.
Review downgraded and escalated alerts across all modules. Log the reason for each routing or severity decision and collect discrepancies for joint calibration with the client. Share recurring noisy signals with the client’s owners and keep their decision on threshold changes in the case history.
Delivery team
Poland, about 15 FTE. 1 security lead, 2 L2 analysts and 12 L1 triage associates.
Systems and tools
Jira Service Management, PagerDuty, Datadog, client SIEM.
Client decision rights
The provider owns incident command, customer notification, and remediation decisions.
Early milestone
The week 3 record shows a 7.2-minute triage time.
Outcomes
Company-reported results, with the baseline, measurement window and supporting record described for each measure.
| Measure | Baseline | Result | Window | Supporting record |
|---|---|---|---|---|
| Manual triage time per alert | 16 min | 2.8 min | 5 months | Jira Service Management |
| False-positive escalation | 38% | 9% | 5 months | PagerDuty report |
| Severity 1 detection to notice | 22 min | 5 min | 5 months | Datadog event report |
| Analyst hours saved | No baseline | 140 / week | 5 months | Timesheet |
Interpreting the results
The five-minute severity-one notification concerns detection to notice. Track acknowledgment and resolution in their own time windows. Read shorter manual triage alongside the false-positive escalation rate and review suspected critical misses. Define analyst hours using recorded effort and the alert population handled.
Recommendations for new clients
For security SaaS and managed security providers, sample escalations and closed alerts for incorrect severity. Define timing from detection to notification separately from mitigation and resolution.
Start with services that have known alert patterns and test cases for overlapping alerts and critical events. Provide the escalation runbook, allowed telemetry, contacts, and approved severity rules. Validate notification routing and evidence quality in a controlled shift before covering additional modules.