Files
windmill/benchmarks/benchmark_suite.ts
pyranota 74b662d8de feat(benchmarks): k8s sim mode + util-group dashboard + reliability fixes
Stand up a minikube-backed simulation subsystem for benching Windmill under
realistic multi-node load, with a per-bench measurement pipeline and a
dashboard renderer that consolidates throughput, queue depth, per-node CPU,
PG latency/conns, OOM events, and per-node CPU-util-vs-oversaturation into
one SVG report.

Sim infrastructure (sim/):
- k8s_provisioner: minikube up + heterogeneous node sizing from topology JSON
- helm_deploy: helm install Windmill with smoke.yaml + local.yaml overlays
- image_cache: pre-load required images so bench bringup is offline-safe
- toxiproxy_k8s: per-node toxiproxy DaemonSet for cross-node latency injection
- cpu_sampler_k8s: privileged DS reading per-cgroup cpu.stat at 10Hz, dual-
  writes to stdout AND a host-mounted log file (/var/log/wm-sim-cpu-sampler/
  sampler.tsv) so heavy benches no longer lose early samples to kubelet log
  rotation
- pg_logging: ALTER SYSTEM + SIGHUP to enable verbose PG logging without restart
- pgbadger: post-bench PG log analysis HTML report
- readiness: pre-bench cluster health check (samplers stable ≥30s, workers
  ready, PG responsive, queue empty, **deploy.status rollout-complete**) —
  the rollout-complete check catches mid-rolling-update fires that previously
  starved m04's sampler under cgroup_mutex contention

Per-bench JSONL pollers, started/finalized alongside the bench loop:
- pod_timeline: 1Hz workers-per-node Ready counts (used for the workers panel)
- oom_poller: live OOM event capture (kernel + kubelet evictions + cgroup)
- pg_latency_poller: 4Hz psql \\timing on SELECT 1 vs kubectl-exec roundtrip
- pg_conn_poller: 1Hz pg_stat_activity by state (active/idle/idle_in_xact)
- node_load_poller: 2Hz /proc/loadavg + /proc/stat procs_running per node

Dashboard renderer (sim/render_report.ts + graph.ts):
- Util group: one panel per node with translucent orange oversaturation area
  BEHIND solid blue CPU-util area, 100% reference line, phase-boundary verticals.
  cols:2 grid wraps after 2 panels per row.
- PG node tinted with [PG] flag in legend across the dashboard.
- Phase-boundary verticals + push-window shaded zones layered consistently.
- All x-axes switched from wall-clock HH:MM to relative seconds-from-bench-
  start. Shared origin sourced from meta.json's bench_start_ms so 0s on every
  panel = the same wall-clock moment (previously each chart picked its own
  earliest sample as origin, causing drift between panels).

Oversaturation metric, with explicit fallback:
- Primary: (procs_running - ncpu) / ncpu × 100 — true CPU run-queue pressure.
- Fallback to load1 when procs_running is missing (older reports).
- load1 overcounted previously because it includes uninterruptible D-state
  procs (PG backends in disk I/O, cgroup_mutex waits), inflating "saturation"
  by 5-10x under load.
- Pure helper extracted to sim/util_metrics.ts; 8 unit tests cover the
  procs_running > load1 preference, the clamp-at-zero, invalid-ncpu cases.

Sampler reliability:
- HostPath log file in addition to stdout so the bench's scp-based collector
  bypasses kubelet log rotation entirely.
- main.ts truncates the host log file on every node before pushers start
  (parallel ssh, best-effort) so it doesn't grow unbounded across runs.
- Collector falls back to kubectl-logs when scp fails for any node.

Workloads (workloads/):
- io_4phase: four-phase IO step (idle → 2.5s → 500ms → 150ms jobs)
- io_150ms_flood / io_300ms_flood / io_1s_flood / io_2s_flood: single-phase
  flood configs to isolate the worker-host CFS context-switch storm vs PG
  contention regime
- burst, ops_day, cpu_*, etc. for other scenarios

Tests:
- sim/util_metrics_test.ts — 8 cases for computeOversatPct
- sim/util_panel_snapshot_test.ts — 5 assertions guarding util-panel SVG
  invariants (orange behind blue, 100% ref line, relative-time ticks NOT
  wall-clock, phase-boundary verticals, shared-origin override)

Helm values:
- sim/values/smoke.yaml — bench-tuned: workers w/ no CPU limit & low mem
  request, PG w/ 3-core request + wm-critical priorityClass + oomImmune +
  maxConnections, app w/ wm-critical + oomImmune + no resource limits.
- sim/values/local.example.yaml — template for the gitignored local.yaml
  that carries the EE license key.
- Depends on the wm-critical PriorityClass + oomImmune + maxConnections
  knobs landing in windmill-helm-charts (separate PR).

graph.ts additions:
- areaFills param: ordered list of per-kind translucent area fills drawn
  before lines, used by the util panel for orange-behind-blue layering
- lineColorOverrides: pin per-kind line colors so oversaturation reliably
  renders orange regardless of d3 ordinal-color insertion order
- highlightKindToken: substring-match flag for the PG-node tint in Node CPU
- xRelativeOriginMs: shared bench-start origin for the relative-time x-axis
- DataPointMulti is now exported for downstream tests

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-06-08 11:43:47 +02:00

199 lines
5.5 KiB
TypeScript

import { Command } from "https://deno.land/x/cliffy@v0.25.7/command/mod.ts";
import { UpgradeCommand } from "https://deno.land/x/cliffy@v0.25.7/command/upgrade/upgrade_command.ts";
import { DenoLandProvider } from "https://deno.land/x/cliffy@v0.25.7/command/upgrade/mod.ts";
import { main as runBenchmark } from "./benchmark_oneoff.ts";
import { VERSION, loadJsonConfig } from "./lib.ts";
// Two accepted on-disk shapes:
// - array of benchmark entries (the original shape; runs unchanged here).
// - object with optional `topology` + `benchmarks`. If `topology` is set the
// suite needs the sim layer, so we redirect; if it's an array-equivalent
// object (no topology) we accept it and run the same as the array form.
// `normaliseConfig` returns the array form so the run loop stays unchanged.
type BenchmarkEntry = {
kind: string;
jobs: number;
noSave?: boolean;
};
type Config = BenchmarkEntry[];
type SuiteFile = Config | {
topology?: string;
benchmarks: BenchmarkEntry[];
};
function normaliseConfig(raw: SuiteFile, configPath: string): Config {
if (Array.isArray(raw)) return raw;
if (raw && typeof raw === "object" && Array.isArray(raw.benchmarks)) {
if (raw.topology) {
console.error(
`[benchmark_suite] Suite ${configPath} declares a topology ` +
`(${raw.topology}). benchmark_suite.ts only runs against an already-` +
`running Windmill; for provisioned topologies use sim/sim.ts instead.`,
);
Deno.exit(2);
}
return raw.benchmarks;
}
throw new Error(
`Suite ${configPath}: expected either an array or an object with ` +
`a "benchmarks" array.`,
);
}
async function warmUp(
host: string,
email: string | undefined,
password: string | undefined,
token: string | undefined,
workspace: string
) {
console.log("%cWarming up...", "font-weight: bold;");
await runBenchmark({
host,
email,
password,
token,
workspace,
kind: "noop",
jobs: 50000,
});
}
async function main({
host,
email,
password,
token,
workspace,
configPath,
workers,
factor
}: {
host: string;
email?: string;
password?: string;
token?: string;
workspace: string;
configPath: string;
workers: number;
factor?: number;
}) {
if (!Deno.args.includes("--no-warm-up")) {
await warmUp(host, email, password, token, workspace);
}
try {
const raw = await loadJsonConfig<SuiteFile>(configPath);
const config = normaliseConfig(raw, configPath);
for (const benchmark of config) {
try {
console.log(
"%cRunning benchmark " + benchmark.kind,
"font-weight: bold;"
);
const result = await runBenchmark({
host,
email,
password,
token,
workspace,
kind: benchmark.kind,
jobs: benchmark.jobs * (factor ?? 1),
});
if (benchmark.noSave) {
continue;
}
if (!result) {
throw new Error("No result returned");
}
const stat = {
value: result.throughput,
ts: Date.now(),
};
let data: (typeof stat)[] = [];
const benchmarkName =
benchmark.kind + (workers > 1 ? `_${workers}workers` : "");
const jsonFilePath = `${benchmarkName}_benchmark.json`;
try {
const remotePath =
"https://raw.githubusercontent.com/windmill-labs/windmill/benchmarks/" +
jsonFilePath;
data = await fetch(remotePath).then((r) => r.json());
} catch (_) {
console.log("No existing data file found, creating new one.");
}
data.push(stat);
await Deno.writeTextFile(jsonFilePath, JSON.stringify(data, null, 4));
} catch (err) {
console.error("Failed to run benchmark", benchmark.kind, err);
}
}
} catch (err) {
console.error(`Failed to read config file ${configPath}: ${err}`);
}
}
await new Command()
.name("wmillbenchsuite")
.description("Run benchmark suite to measure throughput of windmill.")
.version(VERSION)
.option("--host <url:string>", "The windmill host to benchmark.", {
default: "http://127.0.0.1:8000",
})
.option("-e --email <email:string>", "The email to use to login.", {
default: "admin@windmill.dev",
})
.option("-p --password <password:string>", "The password to use to login.", {
default: "changeme",
})
.env(
"WM_TOKEN=<token:string>",
"The token to use when talking to the API server. Preferred over manual login."
)
.option(
"-t --token <token:string>",
"The token to use when talking to the API server. Preferred over manual login."
)
.env(
"WM_WORKSPACE=<workspace:string>",
"The workspace to spawn scripts from."
)
.option(
"-w --workspace <workspace:string>",
"The workspace to spawn scripts from.",
{ default: "admins" }
)
.option("-c --config-path <config:string>", "The path of the config file", {
required: true,
})
.option("--no-warm-up", "Skip the warm up phase.")
.option(
"--workers <workers:number>",
"Number of workers that are used to run the benchmarks (only affect graph title)",
{ default: 1 }
)
.option("--factor <factor:number>", "Factor to multiply the number of jobs by.", {
default: 1,
})
.action(main)
.command(
"upgrade",
new UpgradeCommand({
main: "main.ts",
args: [
"--allow-net",
"--allow-read",
"--allow-write",
"--allow-env",
"--unstable",
],
provider: new DenoLandProvider({ name: "wmillbench" }),
})
)
.parse();