


Spring Boot pplication on AWS Lambda - Part Measuring cold and warm starts with GraalVM Native Image and memory settings
Jan 07, 2025 am 07:17 AMIntroduction
In the article part 12 of our series we explored how to develop and deploy Lambda function with Custom Runtime containing GraalVM Native Image with GraalVM 22 runtime created from Spring Cloud Function AWS application. In the part 13 we measured performance (cold and warm starts) of such a Lambda function with 1024 MB of memory.
In this article, we'll measure the performance (cold and warm starts) of the Lambda function using this approach with different memory settings between 256 and 1536 MBs to explore the trade-off between cost and performance.
Measuring cold and warm starts of Lambda function with Custom Runtime containing GraalVM Native Image with different memory settings
We'll re-use exact the same experiment described in the part 13 of this article series but with different memory settings between 256 and 1536 MBs.
Here are results of the experiment:
Cold (c) and warm (m) start time in ms:
Memory setting | c p50 | c p75 | c p90 | c p99 | c p99.9 | c max | w p50 | w p75 | w p90 | w p99 | w p99.9 | w max |
---|---|---|---|---|---|---|---|---|---|---|---|---|
256 MB | 1634.84 | 1659.54 | 1691.35 | 1778.03 | 1785.15 | 1785.7 | 6.56 | 6.99 | 7.63 | 18.33 | 372.54 | 857.7 |
512 MB | 1244.44 | 1278.48 | 1313.45 | 1414.28 | 1421.36 | 1421.94 | 6.66 | 7.10 | 7.94 | 25.41 | 181.86 | 414.99 |
768 MB | 1111.53 | 1126.07 | 1139.66 | 1192.08 | 1202.86 | 1203.07 | 6.58 | 6.93 | 7.48 | 12.46 | 115.18 | 278.91 |
1024 MB | 1051.03 | 1061.58 | 1080.86 | 1119.34 | 1149.45 | 1230.28 | 6.45 | 6.77 | 7.33 | 12.50 | 90.92 | 218.17 |
1280 MB | 1022.02 | 1035.39 | 1058.41 | 1065.76 | 1104.64 | 1174.79 | 6.58 | 6.96 | 7.54 | 12.37 | 70.77 | 271.13 |
1536 MB | 1009.83 | 1029.20 | 1048.41 | 1161.32 | 1116.24 | 1148.24 | 6.66 | 7.04 | 7.75 | 12.08 | 63.03 | 215.62 |
Conclusion
In this article, measured cold and warms starts of the Lambda function using Custom Runtime containing GraalVM Native Image with GraalVM 21 runtime created from Spring Cloud Function AWS application introduced in the part 12 having different memory settings between 256 and 1536 MBs.
We observe similar things as described in the article Pure Lambda function with GraalVM Native Image - Measuring cold and warm starts using different Lambda memory settings. Warm start times are very close to each other also with the lower Lambda function memory setting like 256 or 512 MBs where the difference is mainly visible for the high percentiles (>= p90). The cold start times are quite high for 256 and 512 MBs and starting from 768 MBs of memory they decrease only a bit by giving Lambda more memory, but without any noticeable difference for memory greater than 1024 MB. Depending on your performance requirements you can give Lambda less memory than 1024 MBs as we initial gave in the sample application and have a very good price performance trade-off with 768 MB or even a bit less memory.
We also shared the same observations as described in the conclusion of the part 13. When we compare cold start times to those measured in the article Pure Lambda function with GraalVM Native Image - Measuring cold and warm starts using different Lambda memory settings (where Lambda function doesn't use any framework like Spring Boot), we see values about 0.5-0.6 seconds lower for each percentile when using pure Lambda function. I personally think that my sample Spring Boot 3 application has some optimization potential as I can't explain such a big difference in the cold start times between those. My (maybe naive) expectation is that the use of the Spring Boot 3 framework with AWS Lambda and GraalVM Native image may only lead to 0.2-0.3 higher cold start times comparing to the usage of the pure Lambda function.
At the time of publishing this article newer versions of frameworks and tools in use became available (GraalVM 23 runtime, Spring Boot 3.4 and version update of Spring Cloud Function library) so you case make the version changes and re-compile GraalVM Native image following the instructions from the part 2 of the series and re-measure the performance. I'll also soon publish the new measurements with these versions and upgrade the example application.
The above is the detailed content of Spring Boot pplication on AWS Lambda - Part Measuring cold and warm starts with GraalVM Native Image and memory settings. For more information, please follow other related articles on the PHP Chinese website!

Hot AI Tools

Undress AI Tool
Undress images for free

Undresser.AI Undress
AI-powered app for creating realistic nude photos

AI Clothes Remover
Online AI tool for removing clothes from photos.

Clothoff.io
AI clothes remover

Video Face Swap
Swap faces in any video effortlessly with our completely free AI face swap tool!

Hot Article

Hot Tools

Notepad++7.3.1
Easy-to-use and free code editor

SublimeText3 Chinese version
Chinese version, very easy to use

Zend Studio 13.0.1
Powerful PHP integrated development environment

Dreamweaver CS6
Visual web development tools

SublimeText3 Mac version
God-level code editing software (SublimeText3)

Hot Topics

There are three main differences between Callable and Runnable in Java. First, the callable method can return the result, suitable for tasks that need to return values, such as Callable; while the run() method of Runnable has no return value, suitable for tasks that do not need to return, such as logging. Second, Callable allows to throw checked exceptions to facilitate error transmission; while Runnable must handle exceptions internally. Third, Runnable can be directly passed to Thread or ExecutorService, while Callable can only be submitted to ExecutorService and returns the Future object to

Java supports asynchronous programming including the use of CompletableFuture, responsive streams (such as ProjectReactor), and virtual threads in Java19. 1.CompletableFuture improves code readability and maintenance through chain calls, and supports task orchestration and exception handling; 2. ProjectReactor provides Mono and Flux types to implement responsive programming, with backpressure mechanism and rich operators; 3. Virtual threads reduce concurrency costs, are suitable for I/O-intensive tasks, and are lighter and easier to expand than traditional platform threads. Each method has applicable scenarios, and appropriate tools should be selected according to your needs and mixed models should be avoided to maintain simplicity

JavaNIO is a new IOAPI introduced by Java 1.4. 1) is aimed at buffers and channels, 2) contains Buffer, Channel and Selector core components, 3) supports non-blocking mode, and 4) handles concurrent connections more efficiently than traditional IO. Its advantages are reflected in: 1) Non-blocking IO reduces thread overhead, 2) Buffer improves data transmission efficiency, 3) Selector realizes multiplexing, and 4) Memory mapping speeds up file reading and writing. Note when using: 1) The flip/clear operation of the Buffer is easy to be confused, 2) Incomplete data needs to be processed manually without blocking, 3) Selector registration must be canceled in time, 4) NIO is not suitable for all scenarios.

In Java, enums are suitable for representing fixed constant sets. Best practices include: 1. Use enum to represent fixed state or options to improve type safety and readability; 2. Add properties and methods to enums to enhance flexibility, such as defining fields, constructors, helper methods, etc.; 3. Use EnumMap and EnumSet to improve performance and type safety because they are more efficient based on arrays; 4. Avoid abuse of enums, such as dynamic values, frequent changes or complex logic scenarios, which should be replaced by other methods. Correct use of enum can improve code quality and reduce errors, but you need to pay attention to its applicable boundaries.

Java's class loading mechanism is implemented through ClassLoader, and its core workflow is divided into three stages: loading, linking and initialization. During the loading phase, ClassLoader dynamically reads the bytecode of the class and creates Class objects; links include verifying the correctness of the class, allocating memory to static variables, and parsing symbol references; initialization performs static code blocks and static variable assignments. Class loading adopts the parent delegation model, and prioritizes the parent class loader to find classes, and try Bootstrap, Extension, and ApplicationClassLoader in turn to ensure that the core class library is safe and avoids duplicate loading. Developers can customize ClassLoader, such as URLClassL

Javaprovidesmultiplesynchronizationtoolsforthreadsafety.1.synchronizedblocksensuremutualexclusionbylockingmethodsorspecificcodesections.2.ReentrantLockoffersadvancedcontrol,includingtryLockandfairnesspolicies.3.Conditionvariablesallowthreadstowaitfor

The key to Java exception handling is to distinguish between checked and unchecked exceptions and use try-catch, finally and logging reasonably. 1. Checked exceptions such as IOException need to be forced to handle, which is suitable for expected external problems; 2. Unchecked exceptions such as NullPointerException are usually caused by program logic errors and are runtime errors; 3. When catching exceptions, they should be specific and clear to avoid general capture of Exception; 4. It is recommended to use try-with-resources to automatically close resources to reduce manual cleaning of code; 5. In exception handling, detailed information should be recorded in combination with log frameworks to facilitate later

HashMap implements key-value pair storage through hash tables in Java, and its core lies in quickly positioning data locations. 1. First use the hashCode() method of the key to generate a hash value and convert it into an array index through bit operations; 2. Different objects may generate the same hash value, resulting in conflicts. At this time, the node is mounted in the form of a linked list. After JDK8, the linked list is too long (default length 8) and it will be converted to a red and black tree to improve efficiency; 3. When using a custom class as a key, the equals() and hashCode() methods must be rewritten; 4. HashMap dynamically expands capacity. When the number of elements exceeds the capacity and multiplies by the load factor (default 0.75), expand and rehash; 5. HashMap is not thread-safe, and Concu should be used in multithreaded
