#!/usr/bin/env python3 """Summarize measured k6 counters and plot reproducible research figures.""" import csv import json from pathlib import Path import sys import matplotlib matplotlib.use("Agg") import matplotlib.pyplot as plt roots = [Path(arg) for arg in sys.argv[1:]] def resource_peaks(root, run): """resource_peaks aggregates five-second samples, including setup and warmup.""" path = root / ("resources-" + run + ".txt") lines = path.read_text().splitlines() pod_snapshot = root / ("pods-" + run + ".txt") if pod_snapshot.exists(): api_pods = {line.split()[0] for line in pod_snapshot.read_text().splitlines() if line.startswith("api-") and "Running" in line} else: api_pods = set() block = set() for line in lines + ["UTC END"]: if line.startswith("UTC "): if block: api_pods = block block = set() elif line.startswith("api-"): block.add(line.split()[0]) peaks = dict(api_cpu_m=0, postgres_cpu_m=0, valkey_cpu_m=0, load_cpu_m=0, api_memory_mib=0, postgres_memory_mib=0, pg_active=0, pg_idle_transaction=0, pg_lock_waiters=0) sample = dict.fromkeys(peaks, 0) for line in lines + ["UTC END"]: if line.startswith("UTC "): for key in peaks: peaks[key] = max(peaks[key], sample[key]) sample = dict.fromkeys(peaks, 0) continue fields = line.split() if len(fields) == 3 and fields[0].startswith(("api-", "postgres-", "valkey-", "load-")): role = fields[0].split("-")[0] if role == "api" and fields[0] not in api_pods: continue if role == "load" and not fields[0].startswith("load-" + run + "-"): continue cpu = float(fields[1][:-1]) if fields[1].endswith("m") else float(fields[1]) * 1000 memory = float(fields[2][:-2]) * (1024 if fields[2].endswith("Gi") else 1) sample[role + "_cpu_m"] += cpu if role in ("api", "postgres"): sample[role + "_memory_mib"] += memory if line.startswith("active|"): sample["pg_active"] += int(line.split("|")[-1]) if line.startswith("idle in transaction|"): sample["pg_idle_transaction"] += int(line.split("|")[-1]) if "|Lock|" in line: sample["pg_lock_waiters"] += int(line.split("|")[-1]) return peaks rows = [] excluded = [] for root in roots: for profile in csv.DictReader((root / "profiles.csv").open()): deployment = root / ("deployment-" + profile["run"] + ".json") if deployment.exists() and json.loads(deployment.read_text())["spec"]["replicas"] != int(profile["replicas"]): excluded.append(str(deployment) + " (replica mismatch)") continue source = root / ("load-" + profile["run"] + ".json") if not source.exists(): excluded.append(str(source)) continue result = json.loads(source.read_text()) metrics = result["summary"]["metrics"] if "message_latency_ms" not in metrics: excluded.append(str(source)) continue def value(name, key="count"): """value reads a measured metric, using zero for an absent error counter.""" return metrics.get(name, {}).get("values", {}).get(key, 0) latency = metrics["message_latency_ms"]["values"] seconds = int(profile["duration"]) operations = value("measured_operations") rows.append(dict(profile, source=str(root), confirmed=value("confirmed_messages"), rps=value("confirmed_messages") / seconds, errors=value("application_errors"), error_percent=100 * value("application_errors") / max(operations, 1), p50_ms=latency["p(50)"], p95_ms=latency["p(95)"], p99_ms=latency["p(99)"], rejected_429=value("application_errors{status:429}"), unavailable_503=value("application_errors{status:503}"), network_errors=value("application_errors{status:0}"), **resource_peaks(root, profile["run"]))) output = roots[0] with (output / "summary.csv").open("w") as stream: writer = csv.DictWriter(stream, fieldnames=list(rows[0])) writer.writeheader() writer.writerows(rows) (output / "excluded.json").write_text(json.dumps(excluded, indent=2)+"\n") fig, axes = plt.subplots(3, 1, figsize=(13, 11), constrained_layout=True) names = [row["profile"] for row in rows] for axis, field, label in zip(axes, ["rps", "p95_ms", "error_percent"], ["Confirmed messages / second", "Write attempt p95, ms", "Failed operations, %"]): axis.bar(names, [row[field] for row in rows], color="#267c75") axis.set_ylabel(label) axis.grid(axis="y", alpha=0.25) axis.tick_params(axis="x", rotation=35) fig.suptitle("Lab 3: measured replica budgets and pool-size experiment") fig.savefig(output / "results.png", dpi=160) fig.savefig(output / "results.svg") plt.close(fig) by_name = {row["profile"]: row for row in rows} if all(f"{series}{n}" in by_name for series in ["fixed", "growing"] for n in [1, 2, 4]): fig, axes = plt.subplots(1, 2, figsize=(11, 4), constrained_layout=True) for series, label in [("fixed", "Fixed total CPU/RAM"), ("growing", "Growing total CPU/RAM")]: for axis, field, ylabel in zip(axes, ["rps", "error_percent"], ["Confirmed messages / second", "Failed operations, %"]): axis.plot([1, 2, 4], [by_name[f"{series}{n}"][field] for n in [1, 2, 4]], marker="o", label=label) axis.set(xlabel="API replicas", ylabel=ylabel, xticks=[1, 2, 4]) axis.grid(alpha=0.25) axis.legend() fig.savefig(output / "matrix.png", dpi=160) fig.savefig(output / "matrix.svg") print(json.dumps(rows, indent=2))