Choose which incident levels send notifications

Understand which severities send alerts, when escalations notify, and how custom checks differ.

Last checked 21 August 2026

TechDash labels open incidents Advisory, Warning, or Critical. The level describes likely impact and affects both notification eligibility and Health Score. It is not a guarantee of delivery because every notification channel also has its own controls.

Understand each result

  • Good means the check passed. Good is not an incident level and does not reduce Health Score. A new Good result can close an open incident. If Notify when a check recovers is On, TechDash can send a recovery for a failure that was eligible to notify.
  • Advisory highlights a lower-impact concern or useful improvement. It may indicate quality, completeness, or best-practice work rather than an immediate outage. Built-in Advisory incidents appear in TechDash but do not send desktop or webhook notifications.
  • Warning identifies a meaningful problem that may affect visitors, crawling, indexing, reporting, or a future renewal. Investigate it soon. Built-in Warning incidents are eligible to notify.
  • Critical identifies the highest likely impact, such as an unavailable page, invalid certificate, or another condition that could seriously disrupt access or search performance. Investigate it first. Built-in Critical incidents are eligible to notify.

Severity describes technical urgency, not the commercial value of the affected page. Review the incident evidence and your business context before deciding what to fix first.

Understand Health Score

Current open incidents affect the website's Health Score. Advisory incidents have weight 1, Warning incidents weight 4, and Critical incidents weight 10. Good results are not open incidents. Resolving or recovering an incident removes it from the open set, while ignoring a check can hide matching findings from the score.

Notification settings do not change severity or Health Score. Turning a channel Off only changes where eligible events are delivered.

Know when channels send

TechDash normally alerts on a state change, not every repeated failing result:

  • A page check can alert when it changes from pass to fail. A first observation may establish a baseline instead.
  • A website-level issue can alert when a new eligible finding appears.
  • An existing issue can alert again when it escalates, such as Warning becoming Critical.
  • An unchanged Warning or Critical result does not repeatedly alert after every scheduled run.

For desktop delivery, Desktop notifications must be On in Notifications, and Desktop must be On under Notify for the affected page. For Slack, Microsoft Teams, or Discord, its webhook field must contain a URL and the corresponding page toggle under Notify must be On. TechDash must be running and able to reach the destination.

Sitewide findings use the monitored / page's channel settings when that page exists. Otherwise, TechDash uses the first monitored page. A matching rule under Notifications > Ignored Checks suppresses the finding and its notifications.

Custom checks and recoveries

For a Custom Check, Severity and Notify are separate. An Advisory custom check can notify when Notify is enabled. A Warning or Critical custom check does not notify when Notify is disabled. Global and page-level channel controls still apply.

Notify when a check recovers controls recovery messages. Turning it Off does not stop a check passing or an incident closing. It only suppresses the external recovery alert.

If an expected notification is missing

  1. Confirm the event was an eligible new failure, escalation, or recovery.
  2. Confirm its built-in severity is Warning or Critical, or its Custom Check has Notify enabled.
  3. Check the global channel setting in Notifications.
  4. Check the affected page's channel under Notify in Monitored Pages.
  5. Review Notifications > Ignored Checks.
  6. Confirm TechDash and monitoring were running.

Use Send test in Notifications to test a channel. A test bypasses severity, page toggles, ignore rules, and recovery rules. The expected result is Test notification sent. in TechDash and a visible test message at the destination.

Related articles