)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":34975,"name":"Jaromír Wysoglad","email":"jwysogla@redhat.com","username":"jwysogla"},"change_message_id":"d23b344e03c7d2c80caf8d7c0960bf6ad26afd6d","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"152f7287_ed0f24b8","updated":"2026-11-09 09:50:21.000000000","message":"I agree with Takashi\u0027s comment, that it\u0027d be nice to have these options exposed to the users as well, since they\u0027re ultimately the ones in control of the webhook endpoints.\n\nBut the bug and the fix make sense and there is a precedence in the rest_notifier_ssl_verify option, which uses a default configured by admin in aodh.conf and could be overwritten by users on alarm creation (that could be the future of these and other options as well).\n\nThank you!","commit_id":"2ddc714ae0e81f4b5d437059c9e009cb4776885c"},{"author":{"_account_id":9816,"name":"Takashi Kajinami","email":"kajinamit@oss.nttdata.com","username":"kajinamit"},"change_message_id":"92849bf840ed9a879c22662ba6d77429d8c66ad3","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"a9b023fd_82e70fe7","updated":"2026-11-09 06:23:41.000000000","message":"I\u0027m still unsure if operators can control redirect behavior about alarms which are owned by users, but we already implement some other options like cert, we could start adding these options there and separately consider exposing some of these to users via alarm options.","commit_id":"2ddc714ae0e81f4b5d437059c9e009cb4776885c"}],"aodh/notifier/rest.py":[{"author":{"_account_id":9816,"name":"Takashi Kajinami","email":"kajinamit@oss.nttdata.com","username":"kajinamit"},"change_message_id":"0e2cf2fa148a3e5c169d2bcac28819d5e9c28826","unresolved":true,"context_lines":[{"line_number":71,"context_line":"                    \u0027Accepted values are 1.2 and 1.3, matching \u0027"},{"line_number":72,"context_line":"                    \u0027keystoneauth1 tls_max_version spelling.\u0027),"},{"line_number":73,"context_line":"    cfg.FloatOpt(\u0027rest_notifier_timeout\u0027,"},{"line_number":74,"context_line":"                 default\u003d10.0,"},{"line_number":75,"context_line":"                 min\u003d0,"},{"line_number":76,"context_line":"                 help\u003d\u0027Timeout in seconds of the connection and of each \u0027"},{"line_number":77,"context_line":"                 \u0027read of the REST notifier. 0 means no timeout. Without \u0027"}],"source_content_type":"text/x-python","patch_set":1,"id":"8dc59c1a_a5a8b4db","line":74,"range":{"start_line":74,"start_character":25,"end_line":74,"end_character":29},"updated":"2026-11-08 05:37:09.000000000","message":"Alarms are defined by users, not admins, and admins have not real control about the behavior. Setting it to low value causes a lot of user impact. I\u0027d rather suggestion we set reasonable large value such as 300 .","commit_id":"3c0abe88d0f2552c0907ae7a48c35913244311c7"},{"author":{"_account_id":39493,"name":"Daniel Gines","display_name":"Daniel Gines","email":"dangines@gmail.com","username":"danielgines"},"change_message_id":"13e0d37b9c1ee3fd76b47a486b8cca614b032f90","unresolved":false,"context_lines":[{"line_number":71,"context_line":"                    \u0027Accepted values are 1.2 and 1.3, matching \u0027"},{"line_number":72,"context_line":"                    \u0027keystoneauth1 tls_max_version spelling.\u0027),"},{"line_number":73,"context_line":"    cfg.FloatOpt(\u0027rest_notifier_timeout\u0027,"},{"line_number":74,"context_line":"                 default\u003d10.0,"},{"line_number":75,"context_line":"                 min\u003d0,"},{"line_number":76,"context_line":"                 help\u003d\u0027Timeout in seconds of the connection and of each \u0027"},{"line_number":77,"context_line":"                 \u0027read of the REST notifier. 0 means no timeout. Without \u0027"}],"source_content_type":"text/x-python","patch_set":1,"id":"2e38ff78_d5d2d0ab","line":74,"in_reply_to":"8dc59c1a_a5a8b4db","updated":"2026-11-08 20:45:07.000000000","message":"Done. The default is now 300 seconds, large enough not to break the existing actions; operators who want a shorter one can lower it, and 0 still disables the timeout.","commit_id":"3c0abe88d0f2552c0907ae7a48c35913244311c7"},{"author":{"_account_id":9816,"name":"Takashi Kajinami","email":"kajinamit@oss.nttdata.com","username":"kajinamit"},"change_message_id":"0e2cf2fa148a3e5c169d2bcac28819d5e9c28826","unresolved":true,"context_lines":[{"line_number":79,"context_line":"                 \u0027the notification of the alarm, and the following \u0027"},{"line_number":80,"context_line":"                 \u0027actions, for minutes.\u0027"},{"line_number":81,"context_line":"                 ),"},{"line_number":82,"context_line":"    cfg.BoolOpt(\u0027rest_notifier_follow_redirects\u0027,"},{"line_number":83,"context_line":"                default\u003dFalse,"},{"line_number":84,"context_line":"                help\u003d\u0027Whether the REST notifier follows HTTP redirections \u0027"},{"line_number":85,"context_line":"                \u0027returned by the alarm action. Redirections let an alarm \u0027"}],"source_content_type":"text/x-python","patch_set":1,"id":"f4a27486_96040613","line":82,"range":{"start_line":82,"start_character":17,"end_line":82,"end_character":47},"updated":"2026-11-08 05:37:09.000000000","message":"I find this problematic because alarms are defined by users. If a user create an alarm which actually requires redirect then it breaks that usage, but they have no way to \"fix\" it unless admin changes this policy. Because users have control about target urls, if they have concern about credential leak about redirects then they can intentionally use the url which may not triggere redirects.","commit_id":"3c0abe88d0f2552c0907ae7a48c35913244311c7"},{"author":{"_account_id":39493,"name":"Daniel Gines","display_name":"Daniel Gines","email":"dangines@gmail.com","username":"danielgines"},"change_message_id":"13e0d37b9c1ee3fd76b47a486b8cca614b032f90","unresolved":false,"context_lines":[{"line_number":79,"context_line":"                 \u0027the notification of the alarm, and the following \u0027"},{"line_number":80,"context_line":"                 \u0027actions, for minutes.\u0027"},{"line_number":81,"context_line":"                 ),"},{"line_number":82,"context_line":"    cfg.BoolOpt(\u0027rest_notifier_follow_redirects\u0027,"},{"line_number":83,"context_line":"                default\u003dFalse,"},{"line_number":84,"context_line":"                help\u003d\u0027Whether the REST notifier follows HTTP redirections \u0027"},{"line_number":85,"context_line":"                \u0027returned by the alarm action. Redirections let an alarm \u0027"}],"source_content_type":"text/x-python","patch_set":1,"id":"95148f97_4dea449c","line":82,"in_reply_to":"f4a27486_96040613","updated":"2026-11-08 20:45:07.000000000","message":"Agreed. rest_notifier_follow_redirects now defaults to True, so the current behaviour is kept. It is only an opt-in for operators who want to make sure that the notifications, and the X-Auth-Token header of the trust+ actions, are only sent to the URL of the alarm action. I updated the release note accordingly.","commit_id":"3c0abe88d0f2552c0907ae7a48c35913244311c7"}],"aodh/tests/unit/test_notifier.py":[{"author":{"_account_id":9816,"name":"Takashi Kajinami","email":"kajinamit@oss.nttdata.com","username":"kajinamit"},"change_message_id":"0e2cf2fa148a3e5c169d2bcac28819d5e9c28826","unresolved":true,"context_lines":[{"line_number":579,"context_line":"        self.client.posts +\u003d 1"},{"line_number":580,"context_line":""},{"line_number":581,"context_line":""},{"line_number":582,"context_line":"class TestRestAlarmNotifierNetwork(tests_base.BaseTestCase):"},{"line_number":583,"context_line":"    \"\"\"Run the REST notifier against local sockets, without mocks.\"\"\""},{"line_number":584,"context_line":""},{"line_number":585,"context_line":"    def setUp(self):"}],"source_content_type":"text/x-python","patch_set":1,"id":"6c643de3_d7857547","line":582,"range":{"start_line":582,"start_character":6,"end_line":582,"end_character":34},"updated":"2026-11-08 05:37:09.000000000","message":"Lanuching internal http server looks overkill to me. If we want this then it should probably be in e2e test, and added to telemetry-tempest-plugin.","commit_id":"3c0abe88d0f2552c0907ae7a48c35913244311c7"},{"author":{"_account_id":39493,"name":"Daniel Gines","display_name":"Daniel Gines","email":"dangines@gmail.com","username":"danielgines"},"change_message_id":"13e0d37b9c1ee3fd76b47a486b8cca614b032f90","unresolved":false,"context_lines":[{"line_number":579,"context_line":"        self.client.posts +\u003d 1"},{"line_number":580,"context_line":""},{"line_number":581,"context_line":""},{"line_number":582,"context_line":"class TestRestAlarmNotifierNetwork(tests_base.BaseTestCase):"},{"line_number":583,"context_line":"    \"\"\"Run the REST notifier against local sockets, without mocks.\"\"\""},{"line_number":584,"context_line":""},{"line_number":585,"context_line":"    def setUp(self):"}],"source_content_type":"text/x-python","patch_set":1,"id":"67e1acc8_3e425af4","line":582,"in_reply_to":"6c643de3_d7857547","updated":"2026-11-08 20:45:07.000000000","message":"Removed. The new tests mock requests.Session.post and check the timeout and allow_redirects arguments built from the options: defaults, custom timeout, 0 (no timeout) and redirects disabled, for http and trust+http actions. If an end-to-end test is wanted, it would indeed fit better in telemetry-tempest-plugin.","commit_id":"3c0abe88d0f2552c0907ae7a48c35913244311c7"}]}