Health notification rules¶
The Rules tab in Health Notifications controls which health conditions create notices, how long the condition must persist, how often notices repeat, and who receives normal and escalated messages.
Platform administrators can maintain global and site rules. Other authorized administrators can view global rules and accessible site rules, but can edit a site rule only when they are assigned to every site included by that rule. Non-platform administrators create site-scoped rules.
Plan the match¶
A rule can narrow notifications by:
- global or selected-site scope;
- target type, or any target type when left blank;
- component type, or the whole target when component criteria are blank;
- optional component-key pattern;
- exact selected statuses, or a minimum severity when no exact statuses are selected.
Component-key patterns support SQL-style % wildcards. For example, camera-% matches keys beginning with camera-. Keep a pattern as narrow as operationally useful; broad overlapping rules can notify the same audience about related conditions.
Available severity values are Unknown, Degraded, Unhealthy, and Offline. Exact status selections take precedence over the minimum-severity fallback.
Configure timing and behavior¶
| Field | Purpose | New-rule default |
|---|---|---|
| Enabled | Allows the rule to participate in notification processing. | Enabled |
| Debounce seconds | Condition must remain present this long before the first notice. | 180 seconds |
| Repeat minutes | Interval between repeat notices while the alert remains active. | 1,440 minutes |
| Escalate after minutes | Makes escalation-level recipients eligible after this duration; zero disables escalation. | 0 |
| Send recovery | Sends a recovery notice when the active condition resolves. | Enabled |
| Suppress during maintenance | Allows an active maintenance window to suppress notices from this rule. | Enabled |
Use debounce to avoid alerting on brief restarts or transient network loss. Set repeat timing according to response urgency and staffed coverage. An extremely short repeat interval can create excessive email or SMS traffic.
Add recipients¶
Choose a recipient type, channel, destination, and escalation level, and then select Add Recipient. The pending recipient must appear in the recipient grid before you save the rule.
| Recipient type | Destination and channel |
|---|---|
| Enter a fixed email destination; channel is Email. | |
| SMS | Enter a fixed SMS destination; channel is SMS. |
| Site Contacts | Uses the target site's configured contacts; choose Email, SMS, or Both. |
Escalation level 0 receives normal notices. Level 1 becomes eligible after Escalate after minutes. Removing a recipient with the trash action is immediate in the editor's pending data but is not persisted until the rule is saved.
The editor does not validate email or phone formatting and can save a rule without a recipient. Verify every destination and ensure that the chosen site-contact channel is populated in the site's contact data. Do not assume an unspecified recipient will be supplied automatically.
Create or edit a rule¶
- Select New Rule, or open an existing editable row.
- Enter a clear, required name that describes the scope and condition.
- Choose Global or Site scope. A site-scoped rule requires at least one site.
- Define target, component, and status matching.
- Set debounce, repeat, escalation, recovery, and maintenance behavior.
- Add and verify every normal and escalation recipient.
- Save the rule.
- Exercise a controlled test condition and verify the alert and delivery records.
Numeric timing fields must contain whole numbers. Blank values may be replaced by service defaults, while nonnumeric input causes the save to fail. The editor does not prevent negative values; use zero only where documented and positive values elsewhere.
Change or retire a rule safely¶
Rule deletion has no confirmation prompt. Prefer disabling a rule first, observing the result, and retaining the configuration until the replacement or retirement is verified. Before changing an active rule, record its current matching, timing, and recipients so the previous behavior can be reconstructed if necessary.
After a change, check:
- Active Alerts for the intended target or component;
- Deliveries for Sent, Failed, or Skipped attempts;
- maintenance-window overlap;
- recipient destination and escalation level;
- other rules that may also match the same condition.
See Troubleshooting when a rule does not fire or creates unexpected noise.