)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":8873,"name":"Assaf Muller","email":"amuller@redhat.com","username":"amuller"},"change_message_id":"9d09ada2aa893b0dd119ca1f4c5c0be7b559433b","unresolved":false,"context_lines":[{"line_number":8,"context_line":""},{"line_number":9,"context_line":"The original patch was based on a hypothesis that fullstack test"},{"line_number":10,"context_line":"failures are caused by random number collisions.  Later observations"},{"line_number":11,"context_line":"revealed the hypothesis isn\u0027t true."},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Change-Id: Idd79b84c291666930d54c78d6e846ad80b30a27b"},{"line_number":14,"context_line":"Related-bug: #1551288"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"ba0121b8_0993ab5c","line":11,"updated":"2016-03-29 13:07:47.000000000","message":"Can you please link to these later observations?","commit_id":"c795d7e68d1aa81bbdadda4134718cc1a2ff8b85"},{"author":{"_account_id":9200,"name":"IWAMOTO Toshihiro","email":"iwamoto@valinux.co.jp","username":"toshii"},"change_message_id":"4a9471482f77be656338d8976bdd57b2d95deb36","unresolved":false,"context_lines":[{"line_number":8,"context_line":""},{"line_number":9,"context_line":"The original patch was based on a hypothesis that fullstack test"},{"line_number":10,"context_line":"failures are caused by random number collisions.  Later observations"},{"line_number":11,"context_line":"revealed the hypothesis isn\u0027t true."},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Change-Id: Idd79b84c291666930d54c78d6e846ad80b30a27b"},{"line_number":14,"context_line":"Related-bug: #1551288"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"ba0121b8_7ca47ca5","line":11,"in_reply_to":"ba0121b8_0993ab5c","updated":"2016-03-30 05:05:45.000000000","message":"I assumed so from comments on the launchpad bug page.\nThis graph suggests the original patch didn\u0027t improve the failure rate (it got merged on Mar 15) but not conclusive because it is concealed by high failure rate of LB tests.\nI\u0027ll try to provide some ground.\n\nhttp://grafana.openstack.org/dashboard/db/neutron-failure-rate?panelId\u003d5\u0026fullscreen","commit_id":"c795d7e68d1aa81bbdadda4134718cc1a2ff8b85"}]}