)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"768a58862240ad541deaa4eec9da04b2eca6fe81","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"284e6b4a_7899a5e3","updated":"2025-09-04 10:00:26.000000000","message":"Agree with @sbauza@redhat.com\u0027s comments. I\u0027ve said this on IRC also, but can\u0027t you just implement this for the VMWare and Ironic drivers but return `None`?","commit_id":"a3c6771d0d79230d1423ee782a28ddbd9c00de2c"},{"author":{"_account_id":26250,"name":"Johannes Kulik","email":"johannes.kulik@sap.com","username":"jkulik"},"change_message_id":"65b863e67bd6f253792693006cc0c0adf5423dd9","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"ed1eae4a_60bbed06","in_reply_to":"284e6b4a_7899a5e3","updated":"2025-09-04 10:06:07.000000000","message":"As said on IRC: why we even found this problem was because `os-hypervisor/detail` is slow. It calls the driver of every hypervisor it returns to get the uptime sequentially. For libvirt, this would be creating a new process `uptime` every time.\n\nSo while we don\u0027t get the noise, we also don\u0027t get a faster call with implementing the method for vmware and ironic. This would only solve half the problem.","commit_id":"a3c6771d0d79230d1423ee782a28ddbd9c00de2c"},{"author":{"_account_id":26250,"name":"Johannes Kulik","email":"johannes.kulik@sap.com","username":"jkulik"},"change_message_id":"529042a4381ffbadb3ae3cb3071a826f452fca9f","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"66a2475b_305b97a3","in_reply_to":"ed1eae4a_60bbed06","updated":"2025-09-04 10:20:08.000000000","message":"I updated the commit-message to more clearly reflect this problem.","commit_id":"a3c6771d0d79230d1423ee782a28ddbd9c00de2c"}],"nova/api/openstack/compute/hypervisors.py":[{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"6f75dcf52279849d4e47a6c7c8d369de999e9ba1","unresolved":true,"context_lines":[{"line_number":102,"context_line":"            # Not all virt drivers support this, and it\u0027s not generally"},{"line_number":103,"context_line":"            # possible to get uptime for a down host"},{"line_number":104,"context_line":"            hyp_dict[\u0027uptime\u0027] \u003d None"},{"line_number":105,"context_line":"            if not CONF.api.disable_hypervisor_uptime_detail:"},{"line_number":106,"context_line":"                try:"},{"line_number":107,"context_line":"                    hyp_dict[\u0027uptime\u0027] \u003d self.host_api.get_host_uptime("},{"line_number":108,"context_line":"                        req.environ[\u0027nova.context\u0027], hypervisor.host)"}],"source_content_type":"text/x-python","patch_set":1,"id":"a5aea3ad_644ab0f5","line":105,"updated":"2025-09-04 09:44:34.000000000","message":"unfortunately, we can\u0027t have API behaviour changed by configuration or we would have interop issues.","commit_id":"a3c6771d0d79230d1423ee782a28ddbd9c00de2c"},{"author":{"_account_id":26250,"name":"Johannes Kulik","email":"johannes.kulik@sap.com","username":"jkulik"},"change_message_id":"4cd4ecf49768730aea0114ac097604180c10b9c8","unresolved":true,"context_lines":[{"line_number":102,"context_line":"            # Not all virt drivers support this, and it\u0027s not generally"},{"line_number":103,"context_line":"            # possible to get uptime for a down host"},{"line_number":104,"context_line":"            hyp_dict[\u0027uptime\u0027] \u003d None"},{"line_number":105,"context_line":"            if not CONF.api.disable_hypervisor_uptime_detail:"},{"line_number":106,"context_line":"                try:"},{"line_number":107,"context_line":"                    hyp_dict[\u0027uptime\u0027] \u003d self.host_api.get_host_uptime("},{"line_number":108,"context_line":"                        req.environ[\u0027nova.context\u0027], hypervisor.host)"}],"source_content_type":"text/x-python","patch_set":1,"id":"21ae301a_4d95338b","line":105,"in_reply_to":"a5aea3ad_644ab0f5","updated":"2025-09-04 09:55:00.000000000","message":"I expected that comment. imho, this does not change the API response in a meaningful way. We already returned `None` whenever a hypervisor was down or the driver didn\u0027t support the `get_host_uptime()` call. This means, clients cannot rely on `uptime` currently already.\n\nThe _api-ref_ states that this field is \"Only reported for active hosts where the virt driver supports this feature.\", which in my eyes is another argument that this does not change expected client behaviour.\n\nStill, we can think about solving this another way e.g. through an async task updating values on the `ComputeNode` object?","commit_id":"a3c6771d0d79230d1423ee782a28ddbd9c00de2c"}]}