Service health
How platform operators interpret service status, verification time, safe failures, impact, and recovery guidance.
On this page
Task and prerequisites
Use workspace connection and run health to identify which capability is affected. You need normal authorized access to that workspace; platform service diagnostics are reserved for allowlisted operators.
Public status guidance does not expose private tenant health or establish live uptime. Have the affected connection or run reference and the observed time available before contacting support.
Status is evidence-bound
Healthy means the corresponding check succeeded at the displayed verification time. Missing credentials, unavailable infrastructure, or stale evidence must remain limited or unavailable.
Operator recovery
Each service entry should identify the affected capability and safe recovery path without exposing secrets or raw provider errors.
Local versus production
Demo and test checks do not establish the health of your live workspace. Check the actual connection, audit or delivery record that supports the task you are performing.
A public example does not establish live provider access or a contractual response target.
Steps and expected result
Inspect the affected connection or audit, then compare its verification time with the period you are reviewing. Check inventory, billing and optional identity/metric sources separately.
Follow the safe recovery guidance through the permitted control and read the next result. The expected result is an evidence-bound status for the affected capability, with partial, stale or unsupported sources still visible.
A healthy connection does not certify that the current report or delivery succeeded.
Limitations and troubleshooting
If health looks current while costs are stale, inspect cost collection rather than rerunning unrelated inventory checks. If the application is unavailable, use the public contact/status guidance with severity, time and safe reference codes.
Do not paste diagnostic dumps, provider secrets or personal data. Contractual response or uptime targets apply only where agreed; this guide is not a production monitoring report or SLA.