release 1.38.0

.gitlab-ci.yml

@@ -12,7 +12,7 @@ before_script:
 
 variables:
   PROJECT: "datakit"
-  CI_VERSION: "1.37.0"
+  CI_VERSION: "1.38.0"
   CI_TESTING_VERSION: $CI_VERSION-testing_$CI_COMMIT_BRANCH
   DISABLED_MR_TARGET_BRANCHES: '/^master|main/'
   DISABLED_MR_TITLE: '/^(\[Draft\]|\(Draft\)|Draft:)/'

internal/export/doc/en/changelog.md

@@ -1,5 +1,46 @@
 # Changelog
 
+## 1.38.0 (September 4, 2024) {#cl-1.38.0}
+
+This release is an iterative update with the following main changes:
+
+### New Features {#cl-1.38.0-new}
+
+- Added Graphite data ingestion (#2337)
+<!-- - Profiling collector now supports real-time metric extraction from profiling files (#2335) -->
+
+### Bug Fixes {#cl-1.38.0-fix}
+
+- Fixed an issue with eBPF network data aggregation (#2395)
+- Resolved a crash issue with the DDTrace telemetry interface (#2387)
+- Addressed a data collection issue with Jaeger UDP binary format (#2375)
+- Resolved an issue with the URL format for data sent by the dial-testing collector (#2374)
+
+### Feature Enhancements {#cl-1.38.0-opt}
+
+- Added collection of multiple fields (`num_cpu/unicast_ip/disk_total/arch`) in the host object (#2362)
+- Other optimizations and fixes (#2376/#2354/#2393)
+
+### Compatibility Adjustments {#cl-1.38.0-brk}
+
+- Adjusted the execution priority of Pipelines (#2386)
+
+    In previous versions, for a specific `source`, such as `nginx`:
+
+    1. If users specified a matching *nginx.p* on the page
+    1. And if users also set a default Pipeline (*default.p*) at the same time
+
+    Then the Nginx logs would **not** be processed by *nginx.p* but by *default.p*. This setting was not reasonable. The adjusted priority is as follows (priority decreasing):
+
+    1. The Pipeline specified for `source` on the observation cloud page
+    1. The Pipeline specified for `source` in the collector
+    1. The `source` value can find the corresponding Pipeline (for example, if the `source` is the log of `my-app`, a *my-app.p* can be found in the Pipeline's storage directory)
+    1. Finally, use *default.p*
+
+    This adjustment ensures that all data can be processed by Pipelines, at least processed by *default.p*.
+
+---
+
 ## 1.37.0 (2024/08/28) {#cl-1.37.0}
 
 This is an iterative release with the following updates:

internal/export/doc/zh/changelog.md

@@ -1,5 +1,46 @@
 # 更新日志
 
+## 1.38.0(2024/09/04) {#cl-1.38.0}
+
+本次发布属于迭代发布，主要有如下更新：
+
+### 新加功能 {#cl-1.38.0-new}
+
+- 新增 Graphite 数据接入（#2337）
+<!-- - Profiling 采集支持从 profiling 文件中实时指标提取（#2335） -->
+
+### 问题修复 {#cl-1.38.0-fix}
+
+- 修复 eBPF 网络数据聚合异常问题（#2395）
+- 修复 DDTrace telemetry 接口崩溃问题（#2387）
+- 修复 Jaeger UDP binary 格式数据采集问题（#2375）
+- 修复拨测采集器数据发送地址格式问题（#2374）
+
+### 功能优化 {#cl-1.38.0-opt}
+
+- 主机对象中，增加多个字段（`num_cpu/unicast_ip/disk_total/arch`）采集（#2362）
+- 其它优化和修复（#2376/#2354/#2393）
+
+### 兼容调整 {#cl-1.38.0-brk}
+
+- 调整 Pipeline 执行优先级（#2386）
+
+    在之前的版本中，对于特定的某个 `source`，比如 `nginx`：
+
+    1. 如果用户在页面上指定了匹配 *nginx.p*
+    1. 如果此时用户还设置了默认 Pipeline（*default.p*）
+
+    那么 Nginx 日志不会通过 *nginx.p* 来切割，而是用 *default.p* 来切割。这个设置是不合理的，调整之后的优先级如下（优先级递减）：
+
+    1. 观测云页面上指定 `source` 对应的 Pipeline
+    1. 在采集器中指定 `source` 对应的 Pipeline
+    1. `source` 取值能找到对应的 Pipeline（比如 `source` 为 `my-app` 的日志，在 Pipeline 对应的存放目录中能找到一个 *my-app.p*）
+    1. 最后再使用 *default.p*
+
+    这样调整之后，能保证所有数据都能被 Pipeline 切割，至少有 *default.p* 来兜底。
+
+---
+
 ## 1.37.0(2024/08/28) {#cl-1.37.0}
 本次发布属于迭代发布，主要有如下更新：
 

internal/plugins/inputs/graphite/cfg.go

@@ -29,6 +29,9 @@ const (
   ## Address to open UDP/TCP, default 9109
   address = ":9109"
 
+  # Whether to open StrictMatch
+  # strict_match = false
+
   ## Example Mapping Configuration
   #[inputs.graphite.metric_mapper]
   # name = "test"
@@ -49,20 +52,17 @@ const (
   # [inputs.graphite.metric_mapper.mappings.labels]
   # job = "${1}_server"
   # outcome = "$3"
-  #provider = "$2"
+  # provider = "$2"
 
   # Regex Mapping Example
-  # [[inputs.graphite_metric_mapper.mappings]]
-  # match = "servers\.(.*)\.networking\.subnetworks\.transmissions\.([a-z0-9-]+)\.(.*)"
+  # [[inputs.graphite.metric_mapper.mappings]]
+  # match = '''servers\.(.*)\.networking\.subnetworks\.transmissions\.([a-z0-9-]+)\.(.*)'''
   # match_type = "regex"
   # name = "servers_networking_transmissions_${3}"
 
   # [inputs.graphite.metric_mapper.mappings.labels]
   # hostname = "${1}"
   # device = "${2}"
-
-  Whether to open StrictMatch
-  # strict_match = false
 `
 )