Relax duration bounds in backpressure prose test

ExportXMLWordPrintableJSON

    • Type: Spec Change
    • Resolution: Unresolved
    • Priority: Unknown
    • None
    • Component/s: Backpressure
    • Needed - No Spec Changes
    • Hide

      Summary of necessary driver changes

      • Update backpressure prose test 5 to subtract a small tolerance (5ms) from the two lower duration bounds in step 10, per the commit below. Drivers MAY adjust the tolerance to fit the timing behavior of their platform. No functional driver changes - test-only.

      Commits for syncing spec/prose tests
      (and/or refer to an existing language POC if needed)

      Context for other referenced/linked tickets

      •  The flaky Node.js CI failure motivated this change.
      Show
      Summary of necessary driver changes Update backpressure prose test 5 to subtract a small tolerance (5ms) from the two lower duration bounds in step 10, per the commit below. Drivers MAY adjust the tolerance to fit the timing behavior of their platform. No functional driver changes - test-only. Commits for syncing spec/prose tests (and/or refer to an existing language POC if needed) https://github.com/mongodb/specifications/commit/9ecb35b1944ade6f7e711316f83c360bf045edfb Context for other referenced/linked tickets  The flaky Node.js CI failure motivated this change.
    • $i18n.getText("admin.common.words.hide")
      Key Status/Resolution FixVersion
      CDRIVER-6465 Done 2.6.0
      CXX-3581 Works as Designed
      CSHARP-6229 Fix Not Needed
      GODRIVER-4141 Fixed 2.10.0
      JAVA-6311 Ready for Work
      NODE-7834 Duplicate
      PYTHON-6092 Fix Not Needed
      PHPLIB-1951 Works as Designed
      RUBY-3957 Works as Designed
      RUST-2501 Won't Fix
      $i18n.getText("admin.common.words.show")
      #scriptField, #scriptField *{ border: 1px solid black; } #scriptField{ border-collapse: collapse; } #scriptField td { text-align: center; /* Center-align text in table cells */ } #scriptField td.key { text-align: left; /* Left-align text in the Key column */ } #scriptField a { text-decoration: none; /* Remove underlines from links */ border: none; /* Remove border from links */ } /* Add green background color to cells with FixVersion */ #scriptField td.hasFixVersion { background-color: #00FF00; /* Green color code */ } #scriptField td.willNotDo { background-color: #FF0000; /* Red color code */ } /* Center-align the first row headers */ #scriptField th { text-align: center; } Key Status/Resolution FixVersion CDRIVER-6465 Done 2.6.0 CXX-3581 Works as Designed CSHARP-6229 Fix Not Needed GODRIVER-4141 Fixed 2.10.0 JAVA-6311 Ready for Work NODE-7834 Duplicate PYTHON-6092 Fix Not Needed PHPLIB-1951 Works as Designed RUBY-3957 Works as Designed RUST-2501 Won't Fix

      Step 10 of prose test 5 ("Overload Errors with baseBackoffMS override base backoff") asserts exact lower bounds on measured durations:

      assertGreaterEqual(exponential_backoff_time, 0.6)
assertGreaterEqual(with_base_backoff_ms_time, 0.3)
assertLess(with_base_backoff_ms_time, 0.6)

      With jitter pinned to 1, the lower bounds equal the exact sum of the backoff sleeps (0.2 + 0.4 = 0.6s, 0.1 + 0.2 = 0.3s), leaving zero tolerance. On some platforms sleep timers run on a millisecond-granularity clock, so a sleep of N ms can complete marginally (sub-millisecond) early when the duration is measured with a higher-resolution clock (e.g. performance.now() in Node.js). With zero tolerance, this fails the assertion.

      Proposal: subtract a small tolerance (e.g. 5ms) from the two lower bounds in step 10.

            Assignee:
            Sergey Zelenov
            Reporter:
            Sergey Zelenov
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

              Created:
              Updated: