Previous Action Items

• [all] track the hanging ingestion job
• [all, future] come up with a solution for retriggers
• [jeads] outline possible corner cases where we need a solution (i.e. thresholds)
  • [jmaher, all] work with perf team to solve this, ideally per page
• [jeads] hack into datazilla: if (osx) branchname += "-Non-PGO"
• [jmaher] post new talos with datazilla friendlier posting and missing 'talos ' in test name
• [all] verify tp5n is accurate
• [all, future] start verifying other data, tsvg*
  • future data to verify would be ts* and counters
• [jhammel] add datazillla links to TBPL
• [all] turn jobs orange for tp5n

Datazilla

• [jeads] The following set of links provide a detailed analysis of the new metrics data over a ~1 month time course

• graphs summarizing data across all test types: tp5n, tsvgr, and tsvgr_opacity
  • http://people.mozilla.com/~jeads/pgo_summary.html
  • http://people.mozilla.com/~jeads/non_pgo_summary.html

• Each of these html pages contains a graph for each tp5n page. Some of them are rather large so you may get an unresponsive script warning, if you have lots of tabs open consider closing a few...
  • http://people.mozilla.com/~jeads/linux_fedora_12_x86_64_page_summary.html
  • http://people.mozilla.com/~jeads/linux_redhat_12_x86_page_summary.html
  • http://people.mozilla.com/~jeads/mac_OSX_10.6.8_x86_64_page_summary.html
  • http://people.mozilla.com/~jeads/mac_OSX_10.7.2_x86_64_page_summary.html
  • http://people.mozilla.com/~jeads/mac_OSX_10.8_x86_64_page_summary.html
  • http://people.mozilla.com/~jeads/win_5.1.2600_x86_page_summary.html
  • http://people.mozilla.com/~jeads/win_6.1.7600_x86_page_summary.html

Talos

• disabled xperf shutdown counters, now xperf is green
• we have datazilla link in tbpl
• minor cleanup for talos tests and issues

Round Table

• BenWa has proposed adding thresholds per test to the tree, everything outside the threshold would be orange and require a perf sherriff to adjust the value. (carry over)
• Consider running only test pages which are stable, so we can turn datazilla on and depend on it
• Consider splitting hardware pool
• Consider running talos on a few builds on a set of machines over and over to see what noise metrics we get (see how much noise is based on differences in machines versus in the build and install of the product)
• Starting a thread with John Daggett who has an idea on a new type/version of tp test. Who wants to be on the thread? Everyone?

Action Items