By "unexpected", I'm referring to service requests sent to oneagent-monitored services that never arrive due to temporary issues. For example, an issue with a proxy system. If the request doesn't get served by the oneagent-monitored service, it gets assigned to a different opaque service.
This has several implications:
- The failure rate on the called service (non-opaque) does not change.
- The parent service may not alert if it continues to respond with a 200 (due to still being available) or if the failures are intermittent.
- Failure rate on the opaque service always often shows a 100% failure rate, even if the problem is very intermittent.
- The opaque service won't belong to the child service's management zone rules unless a specific "public" domain/url-based rule is added. Therefore it may not get picked up by existing alert profiles.
I'm curious if the community has any suggestions based on how they've solved for this.
