Administration & configuration

All otwatch configuration is maintained as code in a version control system and deployed via a pipeline. Changes are reviewed, tested in QA and rolled out to production; every change is traceable with author, time and content.

Alert flow from check to resolution.

Checks and thresholds

Each check has a stable check ID. The ID links the alert rule, the alert, the dashboard panel and the runbook with meaning, likely causes and first actions. The check catalogue and runbooks are available to customers in the otwatch HID portal.

Thresholds are set per check and can differ per environment (e.g. PROD stricter than TEST) and per customer.
A fault must persist for a defined duration (typically 1–2 minutes) before an alert is raised, to avoid false alarms.
Checks can be staged, e.g. Warning at 80 % and Critical at 95 % disk usage.

Severity and priorities

Severity	Priority	Meaning	Notification
Critical	P1 / P2	Production impaired or at immediate risk	Immediately, on-call (push, SMS, phone call), repeated until acknowledged
Warning	P3 / P4	Function impaired, workaround available, or threshold approaching	Ticket and e-mail during service hours
Information	P5 / P6	Notable event without direct impact	Dashboards and reports

Alert routing and channels

Alertmanager routes alerts by severity, environment, component, team and time. Supported channels:

Jira Service Management (tickets, on-call schedules, escalation, acknowledgement)
E-mail and Microsoft Teams
Webhooks / REST to ITSM and event management platforms (e.g. ServiceNow, Jira, Splunk)

Related alerts are grouped and deduplicated; dependency rules suppress follow-up alerts (e.g. Content Server alerts while the database is down), so the root cause is notified first. A watchdog alert reports a failure of the monitoring itself.

Maintenance windows

One-time: a silence with start and end time, created in advance or ad hoc.
Recurring: time intervals by weekday, time and day of month (e.g. every Monday 06:00–08:00).
Scope: a whole environment, a server, a component or single checks.

During a maintenance window, notifications are suppressed but metrics are still recorded. Alerts still active after the window are notified normally. Each window is recorded with author, reason and duration.

Users and roles

Role	Permissions
Viewer	View dashboards, alerts and reports
Editor	Create and change dashboards, explore metrics and logs, manage maintenance windows
Admin	Users, teams, data sources and configuration

Users sign in with single sign-on (Microsoft Entra ID, OTDS or Active Directory); roles are assigned from directory groups. Multi-factor authentication is enforced by the identity provider.

Dashboards and reports

Ready-made dashboards per OpenText component, customisable per customer and role.
Drill-down from the service overview to components, nodes, processes, alerts and logs.
Periodic reports (availability, performance, capacity, SLA) as PDF, Excel or CSV, distributed by e-mail.

Data retention

Metrics and logs are retained for at least 12 months; the retention period is configurable per deployment and log source.