Version | Release date | Description |
4.1.12 | October 10, 2024 | Upstream and downstream interface data errors in certain scenarios are fixed. Network metrics may keep increasing. The bug is fixed. Some metrics may be collected for multiple times. The bug is fixed. Custom parameters extracted in certain scenarios may not take effect. The bug is fixed. Exception instrumentation is added to record all exceptions.
|
4.1.11 | September 15, 2024 | The principle for calculating quantiles is adjusted. The CPU and memory overheads of agents are reduced. A maximum of 10,000 spans can be created in a single call. A policy can be configured to extract monitored thread names in a thread pool. The default policy is REPLACE_ALL_DIGITAL, indicating that all digits are replaced by asterisks (*). You can set the value to REPLACE_TAIL_DIGITAL, indicating that the last character of each monitored thread name is replaced by an asterisk (*) if the character is a digit. JVM loading and unloading metrics are collected. New parameters, such as the quantile, and upstream and downstream services, are added to configure data collection conditions. The memory diagnostics feature may result in system crashes. The bug is fixed. The traces with the code diagnostics feature may not be found in certain scenarios. The bug is fixed. Spans may fail to be reported due to a null pointer of the ons-plugin plug-in. The bug is fixed. Kafka may fail to be initialized. The bug is fixed. Out-of-memory (OOM) errors may occur due to failure to read streaming tables in MongoDB. The bug is fixed. The environment variable ARMS_HOST_TAGS may fail to be parsed. The bug is fixed. Monitoring data may fail to be reported in thread pools. The bug is fixed. Certain vulnerabilities are fixed.
|
4.1.9 | July 10, 2024 | The name of a Serverless Application Engine (SAE) application may fail to be changed by adding the arms.AppName parameter. The bug is fixed. In Lettuce 5.1, spans may fail to be compressed after repeated calls. The bug is fixed. SQL statements may fail to be parsed, and database metrics may fail to be reported. The bugs are fixed. Jedis and Lettuce pool monitoring data may be missing. The bug is fixed. Metric reporting may be disrupted. The bug is fixed. ClassLoader enhancement rarely fails, leading to application startup failures. The bug is fixed. The http.status_code field is added to spans. Crash logs can be automatically collected by using a script.
|
4.1.8 | June 29, 2024 | |
4.1.7 | June 26, 2024 | |
4.1.6 | June 13, 2024 | The JFinal and Webex frameworks are supported. The pull mode of ApsaraMQ for RocketMQ V4.x is supported. Trace IDs and span IDs can be automatically filled in in logs. Base exceptions are supported in the exception settings. Original status codes can be recorded in HTTP metrics. Some connection methods of OceanBase and PolarDB are not supported. The bug is fixed. WebSocket requests may be identified as failed requests. The bug is fixed. Out-of-memory (OOM) issues in some asynchronous scenarios are fixed.
|
4.1.3 | May 27, 2024 | |
4.1.1 | May 17, 2024 | |
4.1.0 | May 15, 2024 | Compatible with open source services, the agent V4.1.0 was built based on the OpenTelemetry agent for Java. All previous features are retained. Custom parameters can be extracted. For more information, see Extract business parameters. PolarDB V2, Cassandra, OceanBase, Apache ShenYu, and Spring Schedule are supported. Automatic asynchronous context pass-through is supported without the need to configure package names. Custom consumers can be configured for RabbitMQ. Custom consumption methods can be configured for Kafka. The priority of protocol-based pass-through can be configured for applications. Custom thread pool monitoring settings can be configured. Metrics of applications deployed in Alibaba Cloud Container Service for Kubernetes (ACK) clusters may be incorrect. The bug is fixed.
|
3.2.10 | July 11, 2024 | |
3.2.9 | June 11, 2024 | |
3.2.8 | May 6, 2024 | Undertow 2.x thread pools can be monitored. The maximum heap memory of a Java virtual machine (JVM) cannot be less than 256 MB. Otherwise, the agent cannot be used. The consumption of ons-plugin V2.x can be monitored. The continuous profiling feature may fail. The bug is fixed. When synchronous servlets are used for instrumentation in Undertow, metric data may be repeatedly calculated. The bug is fixed. Nacos 2.x version conflicts are fixed. Jackson later version conflicts are fixed. Data query based on the /error interface may fail. The bug is fixed. The W3C protocol is not fully incompatible with the agent. The bug is fixed. WebSphere HTTP status codes may fail to be recorded. The bug is fixed. Some Jedis and Redisson features are unavailable. The bug is fixed. The instrumentation of Apache Tomcat V10 may fail due to the Servlet API. The bug is fixed.
|
3.1.4 | January 4, 2024 | JDK 21 is supported. The hotspot code feature can be used with the continuous profiling feature to diagnose slow traces. The logic and success rate of importing data collected by the agent are optimized. The CPU and memory overheads of the agent, and the size of the old generation consumed by objects are reduced. The Vert.x WebServer component V4.X is provided. The instrumentation of the XXL-JOB and OceanBase plug-ins may fail. The bug is fixed.
|
2.9.1.2 | October 12, 2023 | The Spring 6.0 and Spring Boot 3.0 components are supported. The PostgreSQL component V42.3.2 and later is supported. Vert.x V3.3 to V3.9 is supported. Metrics may be missing when you use RedisTemplate to call the Jedis component. The bug is fixed. The asynchronous context of a trace may fail to be passed through when Reactor is used. The bug is fixed. The bug that the instrumentation code of the Kafka client may be overwritten is fixed.
|
2.9.0 | July 13, 2023 | By default, Java Virtual Machine (JVM) system parameters, Kubernetes YAML files, and method input parameters collected from applications can be desensitized. The metrics of the Z Garbage Collector (ZGC) can be displayed. Exceptions caught by the try...catch statement can be excluded from statistics. The bug that host IP addresses may fail to be obtained is fixed. The instrumentation of Reactor Netty is optimized. The number of requests and response time can be accurately recorded.
|
2.8.3 | March 31, 2023 | The following components are supported: SequoiaDB, Informix, ElasticJob, OceanBase, and Db2. The continuous profiling feature is supported in regions outside the Chinese mainland. The memory snapshot feature is replaced with the Alibaba Cloud Application Troubleshooting Platform (ATP) solution. JDK 8, JDK 11, and JDK 17 are supported.
|
2.7.3.5 | July 27, 2022 | JDK 8 and earlier, and JDK 11 are supported by the agent V2.7.3.5. Only JDK 17 is supported by the agent V2.7.3.5-jdk17. The overall performance of the agent is optimized. The bugs of Lambda support in JDK 11 and Lettuce support are fixed.
|
2.7.1.4 | April 26, 2022 | Bean instrumentation of SpringMongoDB, ShardingJDBC, and Dubbo, and tracing for future scenarios are supported. Metrics returned by Lettuce are supported. RASP Application Security is enabled by default. Logs can be collected by using an agent. The CPU utilization of processes can be obtained. Startup analysis, dependency information, and configuration information are supported. The performance of the agent is optimized and the success rate of data reporting is improved. Some known bugs are fixed, such as Spring security vulnerabilities, SQL dimension divergence, and inaccurate sample rate of the Basic Edition.
|
2.7.1.3 | October 18, 2021 | XXL-JOB, SchedulerX, Elasticsearch, and Eclipse Vert.x are supported. Arthas diagnostics, client-side custom sampling, application anti-DDoS protection, and risky component detection are supported. Metrics such as JDBC fetch size and Redis hit ratio are supported. Pool monitoring is optimized. The user experience for some asynchronous scenarios is optimized.
|
2.7.1.2 | September 17, 2021 | Plug-ins such as Kafka, RocketMQ, and Sofa are supported. Thread pool monitoring is supported. The trace information is displayed in the ARMS console. Function Compute scenarios are supported. URL convergence based on Spring annotations is supported. Canary release is supported for Dubbo, Spring Cloud, and Spring Cloud Gateway. Registry migration is supported for Nacos and Eureka. Graceful service degradation is supported. Zone affinity is supported for Dubbo and Spring Cloud. The startup time and memory usage are optimized. Some bugs related to memory snapshots, thread diagnostics, data reporting, and SDKs are fixed. Some security vulnerabilities are fixed.
|
2.7.1.1 | August 14, 2020 | NoSQL monitoring is supported. Routing of microservice tags is supported. Compression for N + 1 invocations is supported. The network connection bug of Alibaba Finance Cloud is fixed, and the memory usage is optimized.
|
2.7.1 | July 16, 2020 | |
2.7.0 | May 20, 2020 | |
2.6.2 | May 20, 2020 | |
2.6.1.2 | March 19, 2020 | |
2.6.1.1 | March 16, 2020 | |
2.6.1 | February 14, 2020 | |
2.6.0.2 | January 2, 2020 | |
2.6.0 | December 17, 2019 | Asynchronous trace is supported. The invocation parameters of Dubbo and HSFProvider are recorded. Some existing plug-in bugs are fixed.
|
2.5.9.5 | November 28, 2019 | The jfinal-undertow plug-in is supported. Some bugs are fixed, such as the failure to obtain Dubbo thread profiling data.
|
2.5.9.3 | November 25, 2019 | Managed Service for OpenTelemetry is integrated with ARMS. Some bugs are fixed, and the agent performance is optimized.
|
2.5.9 | September 6, 2019 | |
2.5.8 | 2019-08-02 | The dual-state alerting feature is supported. Alert rules can be configured for metrics that have only two states: yes and no, or being and not being. The plug-ins of Dameng Database (a database service provided by a Chinese company) are supported.
|
2.5.7.2 | July 30, 2019 | Metrics of JVM Metaspace are supported. HTTP status codes to be ignored can be customized. By default, all HTTP status codes greater than 400 are counted as errors. To ignore specific HTTP status codes greater than 400 so that they are not counted as errors, you can specify them in value. For more information, see Customize application settings.
|
2.5.7 | July 11, 2019 | The dependent Fastjson version that has security vulnerabilities is upgraded. |
2.5.6.1 | June 28, 2019 | Dubbo and MariaDB plug-ins are supported. Bound SQL values can be obtained by custom configurations. The variable values bound to PrepareStatement can be captured. The variable values take effect without the need to restart the application. For more information, see Customize application settings. Memory is optimized and some bugs are fixed. The dependency on Log4j is removed to prevent conflicts.
|
2.5.6 | June 7, 2019 | |
2.5.5 | June 3, 2019 | |
2.5.3 | March 15, 2019 | Thread metrics can be reported when applications are running. The Spring-Data-Redis plug-in is supported. The plug-in of the Druid database connection pool is supported.
|
2.5.2 | February 21, 2019 | The number of file handles can be collected. Instantaneous values for garbage collection (GC) time and for the number of GC operations can be reported. The maximum length of request parameters can be customized. For more information, see Customize application settings.
|
2.5.1 | January 14, 2019 | Trace compression is supported. For more information, see Customize application settings. Application monitoring jobs can be created without the need to use the ARMS console. Features of the agent are optimized, and some bugs are fixed.
|
2.5.0 | December 28, 2018 | The agent can be connected without the need to restart the application. Host monitoring is improved and the Windows operating system is supported. Supported Spring-webflux. Features of the agent are optimized, and some bugs are fixed.
|
2.4.6 | October 26, 2018 | The Google Remote Procedure Call (gRPC), Thrift, and XMemcached plug-ins are supported. Topology views of API calls are supported. Topology views that cover the frontend and backend are supported.
|
2.4.5 | September 17, 2018 | The Lettuce plug-in (JRE version 1.8 or later) is supported. The MongoDB plug-in is supported. Exception details can be captured.
|
2.4.4 | August 6, 2018 | Thread profiling data of applications can be reported. Memcached caching is supported. Exception filtering can be customized. For more information, see Customize application settings.
|
2.4.3.1 | June 29, 2018 | WebLogic servers are supported. Undertow servers are supported. Memory usage of the agent is optimized. The time required to start and load the agent is shortened. The bug that JVM monitoring and host monitoring metrics cannot be reported is fixed.
|
2.4.3 | May 18, 2018 | The monitoring metrics of Message Queue for Apache RocketMQ can be captured. Monitoring methods can be customized. The bug of frequent log output in throttling scenarios is fixed. The maximum length of the method stack can be customized. For more information, see Customize application settings. The sampling feature is optimized. Abnormal traces are excluded.
|
2.4.2 | April 19, 2018 | Custom configuration details can be read. Trace information can be retrieved by using SDKs. JVM metrics such as threads, the number of GC operations, and the duration of GC operations can be collected. HSF calls can be monitored. Host monitoring metrics related to CPU, memory, networks, and disks can be collected. The bug that the ./shutdown.sh process may be stuck in the Tomcat environment is fixed.
|
2.4.1 | March 24, 2018 | JVM monitoring, such as the reporting of heap memory and non-heap memory is supported. Play Framework 1.4.4 is supported. The sampling rate, agent switch, log level, and threshold can be configured. For more information, see Customize application settings.
|
2.4.0 | February 14, 2018 | PostgreSQL databases are supported. ARMS can be connected to Alibaba Cloud Elastic Compute Service (ECS) instances in each region over the internal network. ARMS Application Monitoring is available for commercial use.
|