)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":24245,"name":"Harald Jensås","email":"hjensas@redhat.com","username":"harald.jensas"},"change_message_id":"c19f2f69f18c7ad8bd1d30a8165a781106ab239b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"e52c23ee_1c2a5f31","updated":"2026-09-07 09:30:24.000000000","message":"Looks good to me.","commit_id":"f86b1eacc5fa9737bd8e118048c56bbe5b3128ed"}],"ironic_prometheus_exporter/messaging.py":[{"author":{"_account_id":15519,"name":"Iury Gregory Melo Ferreira","display_name":"Iury Gregory","email":"iurygregory@gmail.com","username":"iurygregory"},"change_message_id":"26573e51d138f0169d5bc8cb70e89be3789a1263","unresolved":true,"context_lines":[{"line_number":107,"context_line":"                if parser is not None:"},{"line_number":108,"context_line":"                    parser(payload, registry)"},{"line_number":109,"context_line":"                else:"},{"line_number":110,"context_line":"                    LOG.debug(\"No sensor parser for event_type %s, only the \""},{"line_number":111,"context_line":"                              \"timestamp metric will be exported.\", event_type)"},{"line_number":112,"context_line":""},{"line_number":113,"context_line":"            # Order of preference is for a node Name, UUID, or"},{"line_number":114,"context_line":"            # payload hostname field to be used (i.e. for conductor"}],"source_content_type":"text/x-python","patch_set":1,"id":"dd7f34c9_0ace8e7d","line":111,"range":{"start_line":110,"start_character":20,"end_line":111,"end_character":79},"updated":"2026-09-01 21:08:28.000000000","message":"In general the patch looks good, I would like to get more clarity regarding this case, where if the hardware has no parser you still save the timestamp metric,\ndo you have a specific use case for this? I\u0027m not sure what value we add by providing only this metric...","commit_id":"f86b1eacc5fa9737bd8e118048c56bbe5b3128ed"},{"author":{"_account_id":290,"name":"Antony Messerli","email":"antony@mes.ser.li","username":"antonym"},"change_message_id":"3377c22cbfabbc91615e57a271970a38ba8f0ba7","unresolved":true,"context_lines":[{"line_number":107,"context_line":"                if parser is not None:"},{"line_number":108,"context_line":"                    parser(payload, registry)"},{"line_number":109,"context_line":"                else:"},{"line_number":110,"context_line":"                    LOG.debug(\"No sensor parser for event_type %s, only the \""},{"line_number":111,"context_line":"                              \"timestamp metric will be exported.\", event_type)"},{"line_number":112,"context_line":""},{"line_number":113,"context_line":"            # Order of preference is for a node Name, UUID, or"},{"line_number":114,"context_line":"            # payload hostname field to be used (i.e. for conductor"}],"source_content_type":"text/x-python","patch_set":1,"id":"95245969_97309407","line":111,"range":{"start_line":110,"start_character":20,"end_line":111,"end_character":79},"in_reply_to":"dd7f34c9_0ace8e7d","updated":"2026-09-01 22:03:29.000000000","message":"The LOG.debug here is new, but the behavior it describes isn\u0027t as the else branch called header.timestamp_registry() unconditionally with no else on the inner chain, so parser-less hardware types like hardware.ilo.metrics already produced a timestamp-only file.\n\nThe metric is labelled with node_uuid/instance_uuid/node_name, so it acts as a per-node freshness signal and it\u0027s the the only indication the conductor to exporter pipeline is still alive.\n\nThat said, I think your point is fair, and I\u0027d be glad to drop it for parser-less types in a follow-up with a release note. I\u0027d just prefer to keep this one scoped to the crash fix since it\u0027s a backport candidate. Let me know what you\u0027d prefer.","commit_id":"f86b1eacc5fa9737bd8e118048c56bbe5b3128ed"}]}