Skip to content
SnelliusSnellius

Health and findings

Every install has one health state: healthy, warning, error or unknown. The overview shows that state and the problems that need attention first, ranked by severity.

Findings

Section titled “Findings”

Each problem is a finding. A finding has:

  • A code, such as KGWUI-013. Codes are stable across releases, so you can search for them and link to them.
  • A severity. An error turns the install red and a warning turns it yellow. Info never changes the state on its own.
  • Evidence: the objects involved, the conditions or metrics that raised it, and since when.
  • A fix: what to change. Some findings also list next-step commands filled in for your install. Snellius never runs them.

“Learn more” on a finding opens its entry in the finding reference in the Snellius repository, for the version you run. The same reference is on this site under Finding codes.

Findings that look the same are grouped on the overview. For example, five routes that point at the same missing Service show as one group with five affected objects.

Where findings come from

Section titled “Where findings come from”
  • Status. Conditions the controller wrote, such as Accepted: False on a route or Programmed: False on a Gateway.
  • Computation. Checks Snellius runs itself, such as a listener attachment that disagrees with the controller (KGWUI-012). These are labelled Computed. The controller’s status always wins; a computed finding is a hint you can check.
  • Metrics and signals. Config the proxies rejected (KGWUI-001), restarting or failing processes (KGWUI-019 to KGWUI-021), version skew (KGWUI-022) and backends without ready endpoints (KGWUI-023).

Unknown is not healthy

Section titled “Unknown is not healthy”

When Snellius cannot see enough to judge, the state is unknown, not healthy. That happens, for example, while a controller restarts or before the first metrics scrape. A check that cannot run, because a chart toggle is off or RBAC denies a read, reads not available and says why. It does not change the install’s state.

Since when

Section titled “Since when”

A finding from a status condition takes the condition’s lastTransitionTime. For other findings, “since” is when Snellius first saw the problem. Snellius keeps that time in memory, so it resets when the UI restarts. With controller logs on, KGWUI-001 can recover its time from the last 15 minutes of logs.