Skip to main content
Axottle’s health system continuously monitors infrastructure: it checks nodes, Flows, gateways, input and transit nodes, WARPs, external proxies, CDNs, and routing rules; it reduces disparate signals to understandable statuses and levels of severity (severity), shows them on the health dashboard, and maintains a tape of incidents.
For whom: panel administrator, SRE, duty engineer. Required rights: panel administrator session.

Monitoring and recovery

Health collects checks and offers action. For Flow, Flow Guard is available: a manual plan with pre-check and confirmation, or a pre-authorized self-recovery policy. By default, modifying actions require manual confirmation. If the operator has enabled the automation of a particular Flow, the server can execute authorized versions of plans and switch the route to explicitly selected reserves. You need fresh data of the running version, confirmation of the failure and compliance with pauses and limits. An AI diagnostic report alone does not allow for infrastructure change; maintaining a user plan does not include it in an automatic policy either. Turning off the automation stops the next steps, but does not roll back the ones already performed. An unconfirmed switching result requires the operator to be checked before restarting.

What exactly is being observed

The checks are organized according to levels, from basic network achievability to end-to-end route verification and external regional samples. For more details, see Levels of Checks.

How to read the state

Each sample returns status; the status of the sample is summarized in the general status of the node, the circuit, and the entire park. healthy, degraded, criticaland unknown (data not yet available) and unsupported Severity incidents are deduced from the general status: critical → error, degraded → warningreturn to healthy → infoThe detailed logic of the summary and the table “state → what it means → what to do” - on the page Conditions and severity.

Section map

Conditions and severity

Sample statuses and summary states, severity, quorum logic according to the connection scheme.

Levels of inspections

That checks each level (behavioral), manual and routine checks.

dashboard health

Summary, tables of nodes and diagrams, incident tape, check button.

Recovery (self-heal)

Event → Identification → Proposed Action → Limitation → Escalation.

Health policies

Verification intervals, AI trigger policy, notification channels, scheduler.

Node suppression

Temporary disabling of notifications and participation of nodes without stopping monitoring.

Runbook.

Step-by-step scenarios for critical allerts: symptom → diagnosis → recovery.
Check the availability of features in the installed version in the panel.