title: VMware vCenter Server Appliance: Update Status
agents: vcsa_health
catalog: app/vmware
license: GPL-2.0-only
distribution: check_mk
description:
 This check monitors available updates for a VMware vCenter Server Appliance
 (VCSA), reports the installed product version, and checks how long ago the
 appliance last queried the update repository. The data is retrieved by the
 special agent {agent_vcsa_health}.

 The state field the appliance exposes on its update endpoint reports a cached
 conclusion and is known to report that the appliance is up to date while
 updates are in fact available. This check therefore uses the pending update
 list as its authoritative source, and falls back to the state field only when
 that list cannot be retrieved.

 Each available update is reported with its version, severity, priority, type,
 release date and whether it requires a reboot. The monitoring state is taken
 from the highest severity present, defaulting to {CRIT} for critical and
 {WARN} for important, moderate, low and unrecognised severities. All of these
 are configurable via the ruleset "VCSA update status".

 An appliance that has stopped querying the repository keeps reporting itself
 up to date, so the age of the last check is alerted separately, defaulting to
 14 days {WARN} and 30 days {CRIT}. Appliances that do not report a last query
 time are not checked for staleness.

 Which source the appliance consults when listing available updates is set in
 the datasource rule. The default reads the cached result of the appliance's
 own last check and performs no network access. Querying the online repository
 is more current but makes the appliance reach out to its update repository on
 every check interval.

 An appliance with nothing pending answers the pending endpoint with an error
 rather than an empty list; this is recognised and reported as no updates
 available rather than as a collection failure.

 Reading the pending list can require a privilege the monitoring account does
 not hold. Where it cannot be read, the check falls back to the appliance's own
 update state field and notes the limitation in its details, rather than
 discarding the check. Because that field can report the appliance as up to
 date while updates are available, the state to report in this situation is
 configurable via the same ruleset.

discovery:
 One service is created per appliance.
