Obervability
Services AudiTrace Resources▾ Contact
Book a free audit
Audit domain · 8 checks

Dynatrace alerting audit: would an incident actually reach a human?

Eight checks on the path between a problem and a person — maintenance windows that never end, anomaly-detection overrides that disable detection per entity, notification integrations pointing at deleted alerting profiles.
Book a free audit All 71 checks

What these checks read

Maintenance windows

Every window, its schedule and its scope. Permanent windows and windows whose end date has passed both suppress alerting indefinitely.

Anomaly-detection overrides

Entity-level settings that turn detection off for a host, service or application. Each one is invisible from the global configuration.

Alerting profiles and notification integrations

Which profile each integration references, and whether that profile still exists. An integration pointing at a deleted profile fails silently.

Problem-to-notification reachability

For each severity, whether any live path exists from a raised problem to a channel a human reads.

Why it matters

Alerting fails in one direction. Nobody notices a suppressed alert; everybody notices a noisy one. So the configuration drifts steadily towards silence: a window opened for a migration and never closed, detection disabled on a chatty host during an incident, an integration rebuilt while the old profile was deleted underneath it.
On the tenant in our case study this domain produced 1,617 maintenance windows that were permanent or expired, and 106 anomaly-detection overrides disabling detection at entity level. Neither number appeared on any dashboard, because a suppressed problem is not a problem — it is nothing at all.
The checks here do not judge your thresholds. They establish whether the wiring is intact: if a problem were raised right now, on this entity, at this severity, is there a live path to a person. Where the answer is no, the finding names the window, the override or the broken integration, with its owner and creation date.
A finding from a real tenant
Permanent or expired maintenance windows
Scale
1,617
Impact
Alerting silently suppressed
Every finding ships with the source behind it and a CSV of the objects it found, so your team can re-run the check without us.

Questions we get about this domain

How can 1,617 maintenance windows be a real number?

They accumulate one at a time. A deployment window is opened per release, per service, and closing it is nobody’s task. Most were created legitimately; the finding is not that they exist but that they never end, and that the alerting they suppress was never re-enabled.

Does the audit change any alerting settings?

No. It reads them and reports. Re-enabling detection on a host that is genuinely noisy without fixing the noise would be worse than the silence, so the remediation order is yours to set — the audit gives you the list, the scope of each item, and what it is currently hiding.

The other audit domains

Nineteen domains, 71 checks, one read-only run. Each domain is a page.
FinOps Where the money goes, and what is recoverable Kubernetes Cluster coverage, versions, recurring failures Alerting Whether an incident would actually reach a human Audit log Who changed what, when, and outside business hours Coverage What is monitored, what only looks monitored RUM Real-user monitoring reach and privacy settings Grail Buckets, retention, annualized storage cost Pipelines Ingest processing, masking, drop rules Synthetic Synthetic monitors that pass, fail, or never run Topology Entity detection quality and entity explosion OpenTelemetry OTel sources and dual instrumentation Security Token scopes, expiry, dormant credentials Workflows Automation that fails or never fires Bizevents Business event provider hygiene Dashboards Dashboard inventory Davis Root-cause attribution rate OpenPipeline Route and pipeline resolution SLO SLO definitions that break on re-detection Tagging Environment tagging convention

See what this domain finds on your own tenant.

One read-only run covers all nineteen domains. You keep the report either way.
Book a free audit See how AudiTrace works

Find out what your Obervability is actually doing.

An audit takes one run. The findings take one meeting. What you do with them is up to you.
Certified Dynatrace experts
+1
Send
Obervability
Dynatrace implementation, managed services and support. Every finding comes with the source behind it.
Services
Services AudiTrace Case study
Guides
Dynatrace audit Dynatrace cost optimization Audit case study Articles
Reach us
© 2026 Obervability · All rights reserved.
Privacy Terms of use Accessibility