Metrics with Micrometer in Spring Boot - Performance & Optimization

Performance: Metrics with Micrometer

MEDIUM IMPACT

This concept affects how efficiently application metrics are collected and reported, impacting monitoring overhead and response time.

Collecting application metrics for monitoring

Spring Boot

MeterRegistry registry = new PrometheusMeterRegistry(PrometheusConfig.DEFAULT);
Counter counter = registry.counter("requests");
for (int i = 0; i < 10000; i++) {
  counter.increment();
}
// Metrics are scraped asynchronously by Prometheus

Using a Prometheus registry enables asynchronous scraping, reducing blocking and CPU overhead during metric updates.

📈 Performance GainNon-blocking metric collection, reducing CPU spikes and improving app responsiveness

Collecting application metrics for monitoring

Spring Boot

MeterRegistry registry = new SimpleMeterRegistry();
Counter counter = registry.counter("requests");
for (int i = 0; i < 10000; i++) {
  counter.increment();
}

Using a simple in-memory registry without asynchronous reporting causes high CPU usage and blocks threads during metric updates.

📉 Performance CostBlocks application threads during increments, increasing response latency under load

Performance Comparison

Pattern	CPU Usage	Memory Usage	Blocking	Verdict
SimpleMeterRegistry with synchronous increments	High CPU during increments	Low memory	Blocks threads on increment	[X] Bad
PrometheusMeterRegistry with async scraping	Low CPU spikes	Moderate memory for buffering	Non-blocking increments	[OK] Good

Rendering Pipeline

Micrometer metrics collection runs alongside application code, affecting CPU and memory usage but not browser rendering directly.

→Application CPU

→Memory Usage

⚠️ BottleneckSynchronous metric updates causing thread blocking

Optimization Tips

1Avoid synchronous metric updates on hot code paths to prevent blocking.

2Use asynchronous or batch reporting registries like PrometheusMeterRegistry.

3Profile CPU and thread usage to detect metric collection overhead.

Performance Quiz - 3 Questions

What is a key performance benefit of using PrometheusMeterRegistry with Micrometer?

A It stores all metrics in memory permanently

B It allows asynchronous metric scraping, reducing blocking in the app

C It disables metric collection to save CPU

D It increases CPU usage to improve accuracy

DevTools: JVisualVM or Java Flight Recorder

How to check: Run your Spring Boot app with Micrometer, profile CPU and thread activity during metric increments, and observe blocking or CPU spikes.

What to look for: Look for thread blocking during metric updates and CPU usage spikes indicating synchronous metric overhead.

Practice

1. What is the primary purpose of Micrometer in a Spring Boot application?

easy

A. To manage database connections efficiently

B. To collect and expose application metrics like counters and timers

C. To handle user authentication and authorization

D. To serve static web content like images and CSS

Solution

1. Step 1: Understand Micrometer's role

   Micrometer is designed to collect metrics such as counters, timers, and gauges to monitor app performance.

2. Step 2: Compare with other options

   Options B, C, and D relate to database, security, and web content, which are not Micrometer's focus.

3. Final Answer:

   To collect and expose application metrics like counters and timers -> Option B

4. Quick Check:

   Micrometer = Metrics collection [OK]

Hint: Micrometer is for metrics, not security or database [OK]

Common Mistakes:

• Confusing Micrometer with security or database tools
• Thinking Micrometer serves static files
• Assuming Micrometer manages app logic

2. Which of the following is the correct way to create a counter metric using Micrometer in Spring Boot?

easy

A. Counter counter = Counter.builder("my.counter").register(meterRegistry);

B. Counter counter = new Counter("my.counter");

C. Counter counter = meterRegistry.createCounter("my.counter");

D. Counter counter = Counter.create("my.counter");

Solution

1. Step 1: Recall Micrometer counter creation syntax

   Micrometer uses builder pattern: Counter.builder(name).register(meterRegistry) to create counters.

2. Step 2: Check other options for correctness

   Options B, C, and D use incorrect constructors or methods not available in Micrometer.

3. Final Answer:

   Counter counter = Counter.builder("my.counter").register(meterRegistry); -> Option A

4. Quick Check:

   Use builder().register() to create counters [OK]

Hint: Use builder() and register() to create metrics [OK]

Common Mistakes:

• Trying to instantiate Counter with new keyword
• Using non-existent createCounter method
• Missing register() call after builder()

3. Given the following code snippet in a Spring Boot app using Micrometer:

Counter counter = Counter.builder("requests.count").register(meterRegistry);
counter.increment();
counter.increment(2);

What will be the value of the counter after these calls?

medium

A. 3

B. 2

C. 0

D. 1

Solution

1. Step 1: Analyze the increment calls

   The first call increments the counter by 1, the second increments by 2.

2. Step 2: Calculate total increments

   Total increments = 1 + 2 = 3.

3. Final Answer:

   3 -> Option A

4. Quick Check:

   1 + 2 = 3 increments [OK]

Hint: Add all increment values to get total count [OK]

Common Mistakes:

• Ignoring the second increment call
• Assuming increment() resets counter
• Confusing increment() with set()

4. Consider this Spring Boot code snippet using Micrometer:

Counter counter = Counter.builder("api.calls").register(meterRegistry);
counter.increment(-1);

What is the issue with this code?

medium

A. meterRegistry is not initialized properly

B. Counter names cannot contain dots

C. Counters cannot be incremented by negative values; this will throw an exception

D. increment() method does not exist on Counter

Solution

1. Step 1: Understand Counter increment rules

   Counters only allow positive increments; negative values cause errors.

2. Step 2: Validate other options

   Counter names with dots are allowed, meterRegistry is assumed valid, and increment() exists.

3. Final Answer:

   Counters cannot be incremented by negative values; this will throw an exception -> Option C

4. Quick Check:

   Counter increments must be positive [OK]

Hint: Counters only accept positive increments [OK]

Common Mistakes:

• Trying to decrement counters with negative increments
• Thinking dots in metric names are invalid
• Assuming increment() method is missing

5. You want to measure the duration of a method execution in a Spring Boot app using Micrometer. Which approach correctly uses a Timer to record the time taken by a method processData()?

hard

A. Timer timer = new Timer("process.time");
timer.start();
processData();
timer.stop();

B. Timer timer = Timer.create("process.time");
timer.record(processData());

C. meterRegistry.timer("process.time").start();
processData();
meterRegistry.timer("process.time").stop();

D. Timer timer = Timer.builder("process.time").register(meterRegistry);
timer.record(() -> processData());

Solution

1. Step 1: Identify correct Timer usage in Micrometer

   Micrometer uses Timer.builder(name).register() and timer.record(Runnable) to measure durations.

2. Step 2: Check other options for errors

   Options A, C, and D use invalid constructors or methods not supported by Micrometer.

3. Final Answer:

   Timer timer = Timer.builder("process.time").register(meterRegistry); timer.record(() -> processData()); -> Option D

4. Quick Check:

   Use Timer.builder().register() and record() for timing [OK]

Hint: Use timer.record(() -> method()) to measure duration [OK]

Common Mistakes:

• Trying to instantiate Timer with new or create()
• Calling start() and stop() manually on Timer
• Passing method call directly instead of lambda to record()