Werk #19983: Stop redfish agent from dropping monitoring data on partial fetch failure

Component	Checks & agents
Title	Stop redfish agent from dropping monitoring data on partial fetch failure
Date	Jun 8, 2026
Level	Trivial Change
Class	Bug Fix
Compatibility	Compatible - no manual interaction needed
Checkmk versions & editions	3.0.0b1 Not yet released Checkmk Community, Checkmk Pro, Checkmk Ultimate, Checkmk Cloud, Checkmk Ultimate MT 2.5.0p7 Not yet released Checkmk Community, Checkmk Pro, Checkmk Ultimate, Checkmk Cloud, Checkmk Ultimate MT 2.4.0p33 Not yet released Checkmk Community, Checkmk Pro, Checkmk Ultimate, Checkmk Cloud, Checkmk Ultimate MT

If a single Redfish endpoint raised an uncaught exception during a poll (for example a timeout, a JSON decoding error, or an unexpected payload from one BMC endpoint), the special agent terminated before emitting the data it had already collected. Healthy responses from earlier endpoints were silently lost.

Visible symptoms on an affected host:

• The Check_MK service reports Missing monitoring data for plugins, redfish_drives, redfish_ethernetinterfaces, redfish_memory, redfish_processors, redfish_storage, redfish_volumes while [special_redfish] still reports Success.
• Previously-discovered services such as Memory Summary or System state ... go UNKNOWN with Item not found in monitoring data.
• The data-source cache for the affected host is noticeably smaller than for comparable hosts.

The agent now emits each section as soon as it has been fully collected, and a failure on one endpoint no longer aborts the rest of the run. The failing endpoint's traceback is written to stderr and the agent continues with the remaining sections.

When --debug is enabled, exceptions still propagate as before so the failing endpoint can be pinpointed during diagnosis.

No user action is required. Affected services repopulate on the next successful agent run.

To the list of all Werks