New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 695533 link

Starred by 1 user

Issue metadata

Status: Archived
Owner:
Last visit > 30 days ago
Closed: Mar 2018
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 3
Type: Bug



Sign in to add a comment

Fix Monarch duration bucket sizes for build timings

Project Member Reported by davidri...@chromium.org, Feb 23 2017

Issue description

The bucket sizes that are being used for build and stage durations is too heavily slanted to very short durations (like RPC call durations).

Given that we have build and stage times that go up to 8+h, it'd be useful to maintain at least 10-15 minute granularity all the way up to 12h instead of having buckets that are longer than an hour at that size.

Here are the current buckets for the build durations:
  Query(Fetch(Raw('monarch.acquisitions.Task', '/chrome/infra/chromeos/cbuildbot/build/durations'), {'metric:build_config': RegexpMatch('master-release')})
        | Window(Delta('1d'))
        | GroupBy(['metric:build_config'], Sum()),
        '1h', '1h')
     |                     [0..1):        0   0%   0%
     |                  [1..1.16):        0   0%   0%
     |             [1.16..1.3456):        0   0%   0%
     |           [1.3456..1.5609):        0   0%   0%
     |          [1.5609..1.81064):        0   0%   0%
     |         [1.81064..2.10034):        0   0%   0%
     |          [2.10034..2.4364):        0   0%   0%
     |          [2.4364..2.82622):        0   0%   0%
     |         [2.82622..3.27841):        0   0%   0%
     |         [3.27841..3.80296):        0   0%   0%
     |         [3.80296..4.41144):        0   0%   0%
     |         [4.41144..5.11726):        0   0%   0%
     |         [5.11726..5.93603):        0   0%   0%
     |         [5.93603..6.88579):        0   0%   0%
     |         [6.88579..7.98752):        0   0%   0%
     |         [7.98752..9.26552):        0   0%   0%
     |          [9.26552..10.748):        0   0%   0%
     |          [10.748..12.4677):        0   0%   0%
     |         [12.4677..14.4625):        0   0%   0%
     |         [14.4625..16.7765):        0   0%   0%
     |         [16.7765..19.4608):        0   0%   0%
     |         [19.4608..22.5745):        0   0%   0%
     |         [22.5745..26.1864):        0   0%   0%
     |         [26.1864..30.3762):        0   0%   0%
     |         [30.3762..35.2364):        0   0%   0%
     |         [35.2364..40.8742):        0   0%   0%
     |         [40.8742..47.4141):        0   0%   0%
     |         [47.4141..55.0004):        0   0%   0%
     |         [55.0004..63.8004):        0   0%   0%
     |         [63.8004..74.0085):        0   0%   0%
     |         [74.0085..85.8499):        0   0%   0%
     |         [85.8499..99.5859):        0   0%   0%
     |          [99.5859..115.52):        0   0%   0%
     |          [115.52..134.003):        0   0%   0%
     |         [134.003..155.443):        0   0%   0%
     |         [155.443..180.314):        0   0%   0%
     |         [180.314..209.164):        0   0%   0%
     |         [209.164..242.631):        0   0%   0%
     |         [242.631..281.452):        0   0%   0%
     |         [281.452..326.484):        0   0%   0%
     |         [326.484..378.721):        0   0%   0%
     |         [378.721..439.317):        0   0%   0%
     |         [439.317..509.607):        0   0%   0%
     |         [509.607..591.144):        0   0%   0%
     |         [591.144..685.727):        0   0%   0%
     |         [685.727..795.444):        0   0%   0%
     |         [795.444..922.715):        0   0%   0%
     |         [922.715..1070.35):        0   0%   0%
     |         [1070.35..1241.61):        0   0%   0%
     |         [1241.61..1440.26):        0   0%   0%
     |          [1440.26..1670.7):        0   0%   0%
     |          [1670.7..1938.02):        0   0%   0%
     |          [1938.02..2248.1):        0   0%   0%
     |          [2248.1..2607.79):        0   0%   0%
     |         [2607.79..3025.04):        0   0%   0%
     |         [3025.04..3509.05):        0   0%   0%
     |          [3509.05..4070.5):        0   0%   0%
     |          [4070.5..4721.78):        0   0%   0%
     |         [4721.78..5477.26):        0   0%   0%
     |         [5477.26..6353.62):        0   0%   0%
     |          [6353.62..7370.2):        0   0%   0%
     |          [7370.2..8549.43):        0   0%   0%
     |         [8549.43..9917.34):        0   0%   0%
     |         [9917.34..11504.1):        0   0%   0%
     |         [11504.1..13344.8):        0   0%   0%
     |         [13344.8..15479.9):        0   0%   0%
     |         [15479.9..17956.7):        0   0%   0%
     |         [17956.7..20829.8):        0   0%   0%
     |         [20829.8..24162.6):        0   0%   0%
     |         [24162.6..28028.6):        0   0%   0%
     |         [28028.6..32513.2):        3 100% 100% #########################
 
We're using SecondsDistribution, a 100-bucket geometric window in ~15% increments (seconds) allows us to cover a wide dynamic range (1 second to 32 days).

Comment 2 by pho...@chromium.org, Aug 11 2017

Status: Fixed (was: Untriaged)

Comment 3 by pho...@chromium.org, Aug 11 2017

Status: Assigned (was: Fixed)
Actually, this isn't fixed, but it's unblocked - we now have a "scale" attribute we could use to adjust this. Also, we should consider increasing the number of buckets we're using.

Comment 4 by pho...@chromium.org, Mar 29 2018

Status: Archived (was: Assigned)

Sign in to add a comment