Incident Ops¶

Incident management and KTLO operations platform: incidents against severity-based SLAs, alerts turned into deduplicated incidents, escalation through an on-call rotation without polling, real-time console, and KTLO analytics with AI-drafted root cause analyses.

This site documents how it is built and how it is operated. The interface contract every module implements is contracts.md.

Live¶

What	URL
Operations console	https://incidents.marceloroman.com.br
Incidents API (Swagger)	https://incidents-api.marceloroman.com.br/swagger
Insights API (OpenAPI)	https://incidents-insights.marceloroman.com.br/docs
Source and local stack	https://github.com/marcelo-roman/incident-ops

System at a glance¶

flowchart LR
    alerts["Prometheus / Alertmanager<br/>Azure Monitor"] -- webhooks --> api
    web["Console<br/>React + TS"] -- REST --> api["Incidents API<br/>.NET 8"]
    web <-- WebSocket --> sig["SignalR"]
    api --> sig
    api --> sql[("Azure SQL")]
    api -- CloudEvents --> bus{{"Service Bus"}}
    bus --> fn["Functions<br/>SLA timers, paging"]
    fn -- escalate --> api
    fn --> logic["Logic App<br/>Teams / Slack / email"]
    ins["Insights<br/>FastAPI + Pandas"] -- export --> api
    ins --> aoai["Azure OpenAI"]
    web -- REST --> ins

Sections¶

Section	Start with
Architecture	overview and C4 diagrams, SLA timers and escalation, deployment and delivery
Operations	support model, alerting, incident response, runbooks
Engineering	quality gates, definition of done
Leadership	stakeholder updates, KTLO vs roadmap
Decisions	architecture decision records

Modules¶

All modules live in one repository, marcelo-roman/incident-ops. Each bounded context is one module (context map).

Module	Bounded context	Role
`services/api`	Incident Management	system of record, SLA clock, alert ingestion, real-time hub
`services/functions`	Escalation	SLA timers, escalation, notifications
`services/insights`	Operational Analytics	KTLO analytics and RCA drafts
`apps/web`	Operations Console	operations console
`infra`		Bicep, Azure DevOps samples, delivery scripts
`docs`		this site
`contracts`		the contract every module implements
`local`		local stack configuration: Prometheus, Alertmanager, Service Bus emulator
`.github/workflows`		one path-filtered workflow per module and a reusable container app deploy