chore(NODE-6871): make tags required and add tags to composite scores (#4480)

Author: nbbeeken

test/benchmarks/driver_bench/src/driver.mts

@@ -8,17 +8,18 @@ const __dirname = import.meta.dirname;
 const require = module.createRequire(__dirname);
 
 export const TAG = {
-  // Special tag that marks a benchmark as a spec-required benchmark
+  /** Special tag that marks a benchmark as a spec-required benchmark */
   spec: 'spec-benchmark',
-  // Special tag that enables our perf monitoring tooling to create alerts when regressions in this
-  // benchmark's performance are detected
+  /** Special tag that enables our perf monitoring tooling to create alerts when regressions in this benchmark's performance are detected */
   alert: 'alerting-benchmark',
-  // Tag marking a benchmark as being related to cursor performance
+  /** Tag marking a benchmark as being related to cursor performance */
   cursor: 'cursor-benchmark',
-  // Tag marking a benchmark as being related to read performance
+  /** Tag marking a benchmark as being related to read performance */
   read: 'read-benchmark',
-  // Tag marking a benchmark as being related to write performance
-  write: 'write-benchmark'
+  /** Tag marking a benchmark as being related to write performance */
+  write: 'write-benchmark',
+  /** A tag for the cpu baseline task */
+  reference: 'reference'
 };
 
 /**
@@ -135,7 +136,7 @@ export type Metric = {
   name: 'megabytes_per_second' | 'normalized_throughput';
   value: number;
   metadata: {
-    tags?: string[];
+    tags: ReadonlyArray<string>;
     improvement_direction: 'up' | 'down';
   };
 };
@@ -150,7 +151,11 @@ export type MetricInfo = {
   metrics: Metric[];
 };
 
-export function metrics(test_name: string, result: number, tags?: string[]): MetricInfo {
+export function metrics(
+  test_name: string,
+  result: number,
+  tags: ReadonlyArray<string>
+): MetricInfo {
   return {
     info: {
       test_name,

test/benchmarks/driver_bench/src/main.mts

@@ -18,7 +18,8 @@ import {
   MONGODB_DRIVER_PATH,
   MONGODB_DRIVER_REVISION,
   MONGODB_DRIVER_VERSION,
-  snakeToCamel
+  snakeToCamel,
+  TAG
 } from './driver.mjs';
 
 const __dirname = import.meta.dirname;
@@ -176,7 +177,7 @@ function calculateCompositeBenchmarks(results: MetricInfo[]) {
     if (compositeName === 'readBench') readBenchResult = compositeAverage;
     if (compositeName === 'writeBench') writeBenchResult = compositeAverage;
 
-    compositeResults.push(metrics(compositeName, compositeAverage));
+    compositeResults.push(metrics(compositeName, compositeAverage, [TAG.spec]));
 
     console.log('avg:', compositeAverage, 'mb/s');
 
@@ -192,7 +193,7 @@ function calculateCompositeBenchmarks(results: MetricInfo[]) {
   console.log('avg:', driverBench, 'mb/s');
   console.groupEnd();
 
-  compositeResults.push(metrics('driverBench', driverBench));
+  compositeResults.push(metrics('driverBench', driverBench, [TAG.spec]));
 
   console.groupEnd();
   return [...results, ...compositeResults];

test/benchmarks/driver_bench/src/runner.mts

@@ -9,19 +9,24 @@ const [, , benchmarkFile] = process.argv;
 
 type BenchmarkModule = {
   taskSize: number;
+  tags: ReadonlyArray<string>;
+
   before?: () => Promise<void>;
   beforeEach?: () => Promise<void>;
   run: () => Promise<void>;
   afterEach?: () => Promise<void>;
   after?: () => Promise<void>;
-  tags?: string[];
 };
 
 const benchmarkName = snakeToCamel(path.basename(benchmarkFile, '.mjs'));
 const benchmark: BenchmarkModule = await import(`./${benchmarkFile}`);
 
 if (typeof benchmark.taskSize !== 'number') throw new Error('missing taskSize');
 if (typeof benchmark.run !== 'function') throw new Error('missing run');
+if (!Array.isArray(benchmark.tags)) throw new Error('tags must be an array');
+if (benchmark.tags.length === 0 || !benchmark.tags.every(t => typeof t === 'string')) {
+  throw new Error('must have more than one tag and all tags must be strings');
+}
 
 /** CRITICAL SECTION: time task took in seconds */
 async function timeTask() {
@@ -81,14 +86,6 @@ function percentileIndex(percentile: number, count: number) {
 const medianExecution = durations[percentileIndex(50, count)];
 const megabytesPerSecond = benchmark.taskSize / medianExecution;
 
-const tags = benchmark.tags;
-if (
-  tags &&
-  (!Array.isArray(tags) || (tags.length > 0 && !tags.every(t => typeof t === 'string')))
-) {
-  throw new Error('If tags is specified, it MUST be an array of strings');
-}
-
 console.log(
   ' '.repeat(3),
   ...['total time:', totalDuration, 'sec,'],
@@ -100,6 +97,6 @@ console.log(
 
 await fs.writeFile(
   `results_${path.basename(benchmarkFile, '.mjs')}.json`,
-  JSON.stringify(metrics(benchmarkName, megabytesPerSecond, tags), undefined, 2) + '\n',
+  JSON.stringify(metrics(benchmarkName, megabytesPerSecond, benchmark.tags), undefined, 2) + '\n',
   'utf8'
 );

test/benchmarks/driver_bench/src/suites/node_specific/cpu_baseline.mts

@@ -1,5 +1,7 @@
 import assert from 'node:assert/strict';
 
+import { TAG } from '../../driver.mjs';
+
 const findPrimesBelow = 1_000_000;
 const expectedPrimes = 78_498;
 
@@ -9,6 +11,9 @@ const expectedPrimes = 78_498;
 
 const stableRegionMean = 42.82;
 export const taskSize = 3.1401000000000003 / stableRegionMean; // ~3MB worth of work scaled down by the mean of the current stable region in CI to bring this value to roughly 1
+
+export const tags = [TAG.reference];
+
 /** @see https://en.wikipedia.org/wiki/Sieve_of_Eratosthenes */
 export function sieveOfEratosthenes(n: number) {
   // Create a boolean array "prime[0..n]" and initialize