External & end-to-end checks
Internal health can be green while the real outcome is broken.
A process can be running, a server reachable and a website responding while the user or recipient still cannot get the result they need.
PMT can receive an independent outside-in or end-to-end result from a checker or receiver you operate, alongside the internal results you already trust.
Component health is not the same as outcome
Mail
The server can run while the message never reaches the recipient.
Website
The site can return HTTP 200 while its enquiry path is broken.
Service
The service can listen locally while nobody outside can reach it.
Look from outside the system.
An external observer you operate does not need server access or shared administrative credentials. It asks the selected external question and reports what it saw to PMT.
For example:
Can this service be reached from this location?
That result can sit beside the internal observation without implying that PMT itself is the external test engine.
End-to-end goes one step further
An end-to-end source or receiver can answer whether the final outcome actually occurred — for example whether a test message, form submission, scheduled export or upload reached the intended destination.
PMT receives the final result; the checker/receiver that performs the test remains outside PMT.
Keep the observations separate.
The usefulness comes from preserving what each observer actually saw rather than collapsing them into an invented whole-system judgement.
INTERNAL
A local service, application or process reports the selected state it knows.
OUTSIDE-IN
An independent checker you operate reports whether the service can be reached or behaves as expected from elsewhere.
FINAL DESTINATION
A receiver or checker you operate confirms whether the selected final outcome arrived.
Internal and external answers can disagree.
That disagreement is useful.
A service may report normal internally while an external observer cannot reach it. A website may be reachable while a final message is not received.
PMT does not collapse those facts into an invented health score.
The operator decides what the disagreement means.
The value is independence, not diagnosis
An external result tells you what the observer saw. It does not automatically tell you why.
A specialist may be more purpose-built for global synthetic monitoring, detailed performance testing, certificate management, mailflow diagnostics or incident escalation.
PMT becomes especially useful when the external result is one part of a broader selected operational picture.
Start at the far end.
Ask: what would prove this is genuinely working for the person at the other end?
Build the smallest check that answers that question.