go.chromium.org/luci@v0.0.0-20240309015107-7cdc2e660f33/analysis/internal/clustering/runs/progress.go (about)

     1  // Copyright 2022 The LUCI Authors.
     2  //
     3  // Licensed under the Apache License, Version 2.0 (the "License");
     4  // you may not use this file except in compliance with the License.
     5  // You may obtain a copy of the License at
     6  //
     7  //      http://www.apache.org/licenses/LICENSE-2.0
     8  //
     9  // Unless required by applicable law or agreed to in writing, software
    10  // distributed under the License is distributed on an "AS IS" BASIS,
    11  // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    12  // See the License for the specific language governing permissions and
    13  // limitations under the License.
    14  
    15  package runs
    16  
    17  import (
    18  	"context"
    19  	"time"
    20  
    21  	"go.chromium.org/luci/server/span"
    22  
    23  	"go.chromium.org/luci/analysis/internal/clustering/shards"
    24  )
    25  
    26  // ReclusteringTarget captures the rules and algorithms a re-clustering run
    27  // is re-clustering to.
    28  type ReclusteringTarget struct {
    29  	// RulesVersion is the rules version the re-clustering run is attempting
    30  	// to achieve.
    31  	RulesVersion time.Time `json:"rulesVersion"`
    32  	// ConfigVersion is the config version the re-clustering run is attempting
    33  	// to achieve.
    34  	ConfigVersion time.Time `json:"configVersion"`
    35  	// AlgorithmsVersion is the algorithms version the re-clustering run is
    36  	// attempting to achieve.
    37  	AlgorithmsVersion int64 `json:"algorithmsVersion"`
    38  }
    39  
    40  // ReclusteringProgress captures the progress re-clustering a
    41  // given LUCI project's test results using specific rules
    42  // versions or algorithms versions.
    43  type ReclusteringProgress struct {
    44  	// ProgressPerMille is the progress of the current re-clustering run,
    45  	// measured in thousandths (per mille).
    46  	ProgressPerMille int `json:"progressPerMille"`
    47  	// Next is the goal of the current re-clustering run. (For which
    48  	// ProgressPerMille is specified.)
    49  	Next ReclusteringTarget `json:"next"`
    50  	// Last is the goal of the last completed re-clustering run.
    51  	Last ReclusteringTarget `json:"last"`
    52  }
    53  
    54  // ReadReclusteringProgress reads the re-clustering progress for
    55  // the given LUCI project.
    56  func ReadReclusteringProgress(ctx context.Context, project string) (*ReclusteringProgress, error) {
    57  	return ReadReclusteringProgressUpTo(ctx, project, MaxAttemptTimestamp)
    58  }
    59  
    60  // ReadReclusteringProgressUpTo reads the re-clustering progress for
    61  // the given LUCI project up to the given reclustering attempt
    62  // timestamp.
    63  //
    64  // For the latest re-clustering progess, pass MaxAttemptTimestamp.
    65  // For re-clustering as it was at a given attempt minute in the
    66  // past, pass the timestamp for that minute. Reading past reclustering
    67  // progress is useful for understanding the versions of rules and
    68  // algorithms reflected in outputs of BigQuery jobs which ran
    69  // using data some minutes old.
    70  //
    71  // When reading historic progress, data for reclustering progress
    72  // as it was part way through a minute is not available; this is
    73  // also why the passed timestamp is called 'upToAttemptTimestamp'
    74  // not an asAtTime.
    75  func ReadReclusteringProgressUpTo(ctx context.Context, project string, upToAttemptTimestamp time.Time) (*ReclusteringProgress, error) {
    76  
    77  	// Reading reclustering progress as at a time in the past can also
    78  	// be achieved with Spanner stale reads, but this does not have
    79  	// good testability. Implement reading past reclustering state
    80  	// using the history we are already storing in the table.
    81  	txn, cancel := span.ReadOnlyTransaction(ctx)
    82  	defer cancel()
    83  
    84  	lastCompleted, err := ReadLastCompleteUpTo(txn, project, upToAttemptTimestamp)
    85  	if err != nil {
    86  		return nil, err
    87  	}
    88  
    89  	last, err := ReadLastUpTo(txn, project, upToAttemptTimestamp)
    90  	if err != nil {
    91  		return nil, err
    92  	}
    93  
    94  	runProgress := 0
    95  
    96  	// For the most recent run, try to read live progress
    97  	// from the reclustering shards table.
    98  	liveProgress, err := shards.ReadProgress(txn, project, last.AttemptTimestamp)
    99  	if err != nil {
   100  		return nil, err
   101  	}
   102  	// Use live progress if it is available.
   103  	if liveProgress.ShardCount > 0 && liveProgress.ShardCount == liveProgress.ShardsReported {
   104  		// Scale run progress to being from 0 to 1000.
   105  		runProgress = int(liveProgress.Progress / liveProgress.ShardCount)
   106  		if runProgress == shards.MaxProgress {
   107  			// If the run is complete.
   108  			lastCompleted = last
   109  		}
   110  	} else {
   111  		// Otherwise, try to use the progress that was on the last
   112  		// run with progress.
   113  		lastWithProgress, err := ReadLastWithProgressUpTo(txn, project, upToAttemptTimestamp)
   114  		if err != nil {
   115  			return nil, err
   116  		}
   117  
   118  		// If the last reclustering run row with progress has the same
   119  		// reclustering objective as the current run, use its progress.
   120  		if last.RulesVersion.Equal(lastWithProgress.RulesVersion) &&
   121  			last.AlgorithmsVersion == lastWithProgress.AlgorithmsVersion &&
   122  			last.ConfigVersion.Equal(lastWithProgress.ConfigVersion) {
   123  			// Scale run progress to being from 0 to 1000.
   124  			runProgress = int(lastWithProgress.Progress / lastWithProgress.ShardCount)
   125  		}
   126  	}
   127  
   128  	return &ReclusteringProgress{
   129  		ProgressPerMille: runProgress,
   130  		Next: ReclusteringTarget{
   131  			RulesVersion:      last.RulesVersion,
   132  			ConfigVersion:     last.ConfigVersion,
   133  			AlgorithmsVersion: last.AlgorithmsVersion,
   134  		},
   135  		Last: ReclusteringTarget{
   136  			RulesVersion:      lastCompleted.RulesVersion,
   137  			ConfigVersion:     lastCompleted.ConfigVersion,
   138  			AlgorithmsVersion: lastCompleted.AlgorithmsVersion,
   139  		},
   140  	}, nil
   141  }
   142  
   143  // IsReclusteringToNewAlgorithms returns whether LUCI Analysis's
   144  // clustering output is being updated to use a newer standard of
   145  // algorithms and is not yet stable. The algorithms version LUCI Analysis
   146  // is re-clustering to is accessible via LatestAlgorithmsVersion.
   147  func (p *ReclusteringProgress) IsReclusteringToNewAlgorithms() bool {
   148  	return p.Last.AlgorithmsVersion < p.Next.AlgorithmsVersion
   149  }
   150  
   151  // IsReclusteringToNewConfig returns whether LUCI Analysis's
   152  // clustering output is in the process of being updated to a later
   153  // configuration standard and is not yet stable.
   154  // The configuration version LUCI Analysis is re-clustering to is
   155  // accessible via LatestConfigVersion.
   156  // Clients using re-clustering output should verify they are using
   157  // the configuration version defined by LatestConfigVersion when
   158  // interpreting the output.
   159  func (p *ReclusteringProgress) IsReclusteringToNewConfig() bool {
   160  	return p.Last.ConfigVersion.Before(p.Next.ConfigVersion)
   161  }
   162  
   163  // IncorporatesRulesVersion returns returns whether LUCI Analysis
   164  // clustering output incorporates all rule changes up to
   165  // the given predicate last updated time. Later changes
   166  // may also be included, in full or in part.
   167  func (p *ReclusteringProgress) IncorporatesRulesVersion(rulesVersion time.Time) bool {
   168  	return !rulesVersion.After(p.Last.RulesVersion)
   169  }