Choosing Platform Monitoring Signals
Key Takeaways
Use the provider status page for shared-service impact and tenant signals for local execution.
Track source health, job timing and results, and provisioning outcomes with different evidence.
Health emails have observation conditions. Use task or tenant monitoring for tighter detection needs.
Verify that a recovered component processes the required workload and clears any backlog.
Monitor layers that can fail independently
For SailPoint operations, pick a user outcome and identify its dependencies. An HR feed must arrive, an aggregation must read it, identities must process changes, and the expected accounts or access may then need updating. A healthy connection is useful evidence for one layer. It provides limited information about whether today's feed included all expected records or whether a subsequent operation completed. Choose signals with an owner and an expected cadence. Compare against a normal baseline before interpreting a smaller account count or slower processing as an outage.
| Layer | Where to monitor | What action follows a changed signal |
|---|---|---|
| SailPoint service and region | SailPoint Status incidents and notifications | Compare incident time and region against affected transactions |
| Tenant component state | Admin Dashboard System Components Status | Inspect the unhealthy source, virtual appliance (VA) or cluster |
| Active and historical jobs | Dashboard Monitor and System Activity, plus source-specific aggregation history | Compare start time, duration, status and covered object counts |
| Processing or control outcomes | Search events, Account Activity or workflow executions | Identify affected operations, technical result and unresolved effect |
| Local VA resources | VA health and supported observability exports | Identify sustained memory, storage or connectivity issues and affected sources |
On the dashboard, recent activity and active jobs show whether the tenant has work in progress. For an aggregation, compare its history timestamps, objects scanned and result to the feed schedule. Agree on how late an import can be before its downstream processes miss their deadline. Also alert on a successful run that scans fewer records than expected when the business impact warrants it. Document the expected format and generation time of the feed so a target-side issue can be distinguished from a tenant scheduling problem. Monitor a provisioning write with its account-level operation result when success on a specific live account is material.
Choose alert routing and timing
SailPoint documents component health notification settings for sources and virtual appliances. Enable the relevant components and choose either the defined admin audience or designated approved email addresses. SailPoint recommends enabling these notifications after deployment and after aggregations are operating as expected. Verify that the recipients include the team that can inspect the affected dependency. Avoid assigning every alert to a shared inbox unless clear ownership and response timing are established.
The documented source health check for email runs once per hour. At the check, a source must have been unhealthy for at least 15 minutes for the health message to be sent. An individual VA communication alert also has an hourly check and a 15-minute observation condition. The documentation describes a recovery notification after at least five healthy minutes at the applicable check. Cluster resource alerts follow a separate evaluation of sustained metrics. These delays matter: an organization with an internal ten-minute requirement cannot use the source health email alone to demonstrate it meets that requirement. Use suitable authorized monitoring of task timing and the necessary tenant or native signals for that deadline. Keep the documented polling conditions separate from a guarantee about exact arrival time in the mailbox.
Work a feed delay and recovery
Suppose the regional status is normal and the source health tile has no warning at 06:30, but today's scheduled HR run has no completion and the service desk reports stale department values. Compare the feed's timestamp with System Activity and the source aggregation record. If the job did not start, inspect its schedule and required input. If it is running or failed, inspect the task and source details. If it completed, compare objects scanned against the feed and inspect relevant identity state and Search timing before concluding that all current records propagated.
After the feed is restored, verify that a new or resumed run processed the current file, that representative identities reflect the corrected data, and that pending effects now completed. Record what alerted, how quickly the team could act, and what established recovery. A normalized service tile closes the health symptom. Reconciliation is the evidence that the affected processing objective has been restored.
See SailPoint's status service, dashboard and activity monitoring, health notification conditions, aggregation history, and VA observability.
An operations team needs to know within ten minutes if a daily feed has failed to reach its processing milestone. Which approach meets the evidence requirement?
Treat the absence of a source health email after ten minutes as a completed feed.
Use the account status from yesterday to establish today’s result.
Track the current feed and job deadline directly with suitable monitoring, and keep source health notifications as a separate signal.
Subscribe only to regional public status pages and ignore tenant jobs.
A source warning clears after an HR feed incident. Which follow-up establishes operational recovery for the missed feed?
Confirm the current aggregation processed the intended input and verify representative identity changes and any pending effects.
Record only the green source status.
Assume clearing the warning also reran every failed or missing job.
Ignore the feed generation time and compare only the account total from the last successful run.
Sections you finish are checked off in the contents.