Glossary
The terms used across Excubius and this documentation, in alphabetical order.
| Term | Definition |
|---|---|
| Accepted risk | A finding an editor decided to live with for a while, with a reason and an end date. It leaves the open list and the security score until the acceptance ends or its severity rises. |
| Alert | A condition that needs attention, raised when a check crosses a threshold, fails to run, or a source reports one. Each alert has a severity, a status and a timeline. |
| Alert inbox | The screen where people see, acknowledge, silence, assign and resolve alerts. By default it shows the active alerts you have not seen yet, most severe first. |
| Availability group | A SQL Server Always On availability group, monitored as one resource with its replicas, databases and listener. Its checks connect through the listener, so they follow the primary. |
| Baseline | The default threshold of a check, as the catalog declares it, with its published source: Microsoft documentation, Azure Monitor recommendations, the CIS Benchmark, vendor documentation, community practice or your recovery objectives. |
| Cannot reach | The one critical alert a resource gets after repeated connection, sign-in or timeout failures, 3 by default. It becomes the root cause of the resource's other alerts. |
| Catalog | The list of everything Excubius can check, kept as YAML files in the product. Every check, its query, baseline and suggested action comes from it. |
| Check | One measurement of one resource on an interval, such as the CPU of a SQL Server instance or the expiry of a certificate. Its result is compared with its thresholds. |
| Collector | The part of Excubius that reaches your systems over read-only connections, runs the checks and reads the sources. It runs on the Excubius server. |
| Credential | A read-only login the collector uses, such as a SQL login or an SSH key, stored on an entity. Its secret is write-only: encrypted and never shown again. |
| Criticality | A number from 1 (low) to 5 (critical) on a resource. It weighs the resource in the security score and helps decide what to work on first. |
| Discovery | A job that reads a SQL Server instance and adds its availability groups, databases and linked servers as child resources. What disappears is marked missing, never deleted. |
| Entity | A node of the tree that organises everything in Excubius: a company, a client, an environment or a team. Every resource lives in one entity, and settings flow down the tree. |
| Error kind | The reason a check could not run, such as auth, unreachable or query. It narrows the cause and decides whether the failure counts towards Cannot reach. |
| Feature switch | An instance-wide on or off setting for an optional capability, such as a module, a source kind or the status page. A super administrator sets it in Settings › Instance. |
| Finding | An alert of the security modules: a posture problem with its evidence, the baseline it is measured against and a suggested action. Findings lower the security score instead of a resource's status. |
| Fingerprint | The identity of an alert's condition, made from the resource, the check and its labels. Repeats of the same condition update one alert instead of opening new ones. |
for duration | How long a value must stay beyond a threshold before the alert opens, such as for 5m. It keeps short spikes from raising alerts. |
| Ingest mode | How a source's data reaches Excubius: Polling, Webhook only or Hybrid, which uses webhooks for alerts and polling for the rest. |
| Inheritance | The rule that a setting made on an entity applies to everything below it, until a lower entity or a single resource overrides it. Thresholds, channels, maintenance windows and roles follow it. |
| Maintenance window | A period in which resources are expected to misbehave. Checks keep running and alerts are recorded, but nobody is notified until it ends. |
| Module | A group of catalog checks for one kind of system or concern, such as sqlserver, linux or security_cloud. Each module is switched on by a feature. |
| Override | A threshold set on an entity or a resource that replaces the inherited value. The screens show where the value in force comes from. |
| Preset | A named set of feature switches and branding an instance starts from, such as enterprise, databases or core. |
| Recommendation | A lower-severity finding, such as a medium Defender for Cloud assessment. It is never notified and counts in the score only while recommendations are shown. |
| Resource | One thing Excubius monitors: an instance, a database, a host, a web endpoint, an Azure resource. It has a type, attributes, tags and checks. |
| Resource type | The kind of a resource, such as SQL Server instance or Web endpoint / portal. It decides the details the resource needs and its default checks. |
| Role | What a person can do on an entity and everything below it: viewer, operator, editor or entity admin. The highest role on the path wins. |
| Root cause | An alert that explains others. While it is open, the alerts that depend on it link to it and their notifications wait. |
| Security score | A number from 0 to 100 per area (identity, database, cloud) and overall, lowered by open findings according to their severity, age and the criticality of the resource. |
| Severity | How serious an alert is: Info, Warning or Critical. Channels can set a minimum severity. |
| Silence | A pause in the notifications of one existing alert for a while. Unlike a maintenance window, it does not cover alerts that do not exist yet. |
| Source | A platform Excubius reads from: New Relic, NinjaOne, Azure Monitor or the native collector. Every source only reads. |
| Status page component | A part of a public status page with a plain-language name, whose status follows resources or single checks, or a manual override. |
| Suggested action | Text from the catalog that tells a person what to do about an alert or a finding. Excubius never runs it. |
| Synthetic check | A check that exercises a real path the way a user would, such as loading a web page, resolving a listener name or rendering a report. |
| Tag | A key:value label on a resource, such as env:prod or client:ciguapa. Tags group resources across the entity tree. |
| Threshold | The warning or critical level a check's value is compared with. It starts from the baseline and can be overridden on entities and resources. |
| Transport | The way the collector talks to one kind of system, such as T-SQL, HTTP, DNS, WinRM or SSH. |
| Webhook URL | The address a source posts its alerts to. It carries a secret token, is shown only once and can be rotated. |