Platform Advisor
The Platform Advisor is the list of active problems with the environment. A background watchdog checks the environment every few minutes and raises an advisory for each active problem. A check that passes again removes its advisory, so the list is always the current state, never a history.
Open Monitoring, then Platform Advisor (/admin/monitoring/advisories). The Monitoring overview banner links here too.
Reading the list
Tiles at the top count advisories by severity: Critical, High, Medium and Low. Filter chips narrow the list to one category: Security, Performance, Availability or Updates.
Each advisory names the problem, shows when it was last checked and first detected, and offers a suggested action. Expand a row for its check code, any compliance tags and the underlying metric.
| Suggested action | What to do |
|---|---|
| Configure | Change a platform setting. The button deep-links to Configuration |
| Upgrade | Update a mismatched image or component |
| Investigate | Open the Cluster or Logs pages |
| Review | Read the detail and decide |
Work the Critical and High rows first, then use the action button to reach the page that fixes each one.
What the checks cover
The checks shipped today fall into four categories.
| Category | Examples |
|---|---|
| Security | Core API in Local authentication mode while the environment looks internet-connected, a TLS certificate expiring, unexpected internet egress from an air-gapped environment, pods with host access |
| Availability | The ingestion queue, object store or database not running, crash-looping or failed pods, a node not ready or under pressure, a growing task backlog, stalled retention enforcement, the object store volume filling up |
| Performance | Recent out-of-memory kills, node count or capacity below the recommended specification, recent Core API, HEAT Auth or v2 API exceptions, high service latency, recent task failures, runner limits too large for the nodes, active templates with no retention policy |
| Updates | Platform deployments running a different image tag from the Core API |
Some advisories carry compliance tags, for example a control reference. They are pointers for your own review, not a compliance claim.
On a single-node appliance the watchdog reports the single-node cluster as a standing Availability advisory, since one node failure takes HEAT offline. This is inherent to the deployment, not a fault to clear.