obs: mirror log attributes into loki lines (#5776)

Co-authored-by: heihutu <heihutu@gmail.com>

.docker/observability/README.md

@@ -170,6 +170,10 @@ Important behavior notes:
 
 - Logs and metrics usually appear during startup, so seeing those two signals
   first is expected.
+- The OpenTelemetry bridge sends `tracing` fields as log attributes. Loki stores
+  those attributes as structured metadata, and the Collector also mirrors the
+  common troubleshooting fields into the log line so simple line filters can
+  find them.
 - Visible trace data usually requires real HTTP/S3/gRPC request traffic after
   startup, because request-path spans are created on demand.
 - `RUSTFS_OBS_LOGGER_LEVEL=info` keeps the top-level request span but filters
@@ -195,6 +199,17 @@ curl -I http://127.0.0.1:9000/health/ready
 # Jaeger:  http://localhost:16686
 ```
 
+For a structured RustFS log such as an inter-node RPC authentication failure,
+the Loki line now includes fields such as `event`, `component`, `subsystem`,
+`failure_reason`, `rpc_service`, `rpc_method`, and `expected_audience`. Useful
+LogQL checks:
+
+```logql
+{service_name="RustFS"} |= "RPC signature verification failed"
+{service_name="RustFS"} |= "failure_reason="
+{service_name="RustFS"} | failure_reason != ""
+```
+
 If logs and metrics are present but traces are sparse, the most common cause is
 "no real request traffic yet" or "`info` level filtered nested spans", not an
 OTLP routing failure.

.docker/observability/README_ZH.md

@@ -169,6 +169,7 @@ RustFS 会自动在该基础 URL 后补全：
 需要注意：
 
 - 启动阶段通常会先看到日志和指标，因此“先有日志/指标、后有 trace”是正常现象。
+- OpenTelemetry bridge 会把 `tracing` 字段作为日志 attributes 发送。Loki 会将这些 attributes 存为 structured metadata，同时 Collector 会把常用排障字段镜像进日志行，方便用简单的行内容过滤直接查到。
 - 可见的 trace 数据通常依赖启动后的真实 HTTP/S3/gRPC 请求流量，因为请求路径上的 span 是按需创建的。
 - `RUSTFS_OBS_LOGGER_LEVEL=info` 会保留顶层请求 span，但会过滤掉很多 `debug` 级别的嵌套 span。
   如果 Tempo 或 Jaeger 中的 trace 看起来很稀疏，建议先改成 `RUSTFS_OBS_LOGGER_LEVEL=debug`，再判断是否是 collector 或 Tempo 问题。
@@ -192,6 +193,14 @@ curl -I http://127.0.0.1:9000/health/ready
 # Jaeger:  http://localhost:16686
 ```
 
+对于 RustFS 结构化日志，例如节点间 RPC 鉴权失败，Loki 日志行现在会包含 `event`、`component`、`subsystem`、`failure_reason`、`rpc_service`、`rpc_method`、`expected_audience` 等字段。常用 LogQL 检查：
+
+```logql
+{service_name="RustFS"} |= "RPC signature verification failed"
+{service_name="RustFS"} |= "failure_reason="
+{service_name="RustFS"} | failure_reason != ""
+```
+
 如果日志和指标已经正常，但 trace 仍然稀疏，最常见的原因通常是
 “还没有真实请求流量”或“`info` 级别过滤了嵌套 span”，而不是 OTLP 路由失败。
 

.docker/observability/otel-collector-config.yaml

@@ -29,11 +29,27 @@ processors:
     limit_mib: 1024
     spike_limit_mib: 256
   transform/logs:
+    error_mode: ignore
     log_statements:
       - context: log
         statements:
-          - set(attributes["message"], body.string)
-          - set(attributes["log.body"], body.string)
+          - set(attributes["message"], body.string) where IsString(body)
+          - set(attributes["log.body"], body.string) where IsString(body)
+          - set(body, Concat([body, " event=", attributes["event"]], "")) where IsString(body) and attributes["event"] != nil
+          - set(body, Concat([body, " component=", attributes["component"]], "")) where IsString(body) and attributes["component"] != nil
+          - set(body, Concat([body, " subsystem=", attributes["subsystem"]], "")) where IsString(body) and attributes["subsystem"] != nil
+          - set(body, Concat([body, " state=", attributes["state"]], "")) where IsString(body) and attributes["state"] != nil
+          - set(body, Concat([body, " result=", attributes["result"]], "")) where IsString(body) and attributes["result"] != nil
+          - set(body, Concat([body, " reason=", attributes["reason"]], "")) where IsString(body) and attributes["reason"] != nil
+          - set(body, Concat([body, " failure_reason=", attributes["failure_reason"]], "")) where IsString(body) and attributes["failure_reason"] != nil
+          - set(body, Concat([body, " rpc_path=", attributes["rpc_path"]], "")) where IsString(body) and attributes["rpc_path"] != nil
+          - set(body, Concat([body, " rpc_service=", attributes["rpc_service"]], "")) where IsString(body) and attributes["rpc_service"] != nil
+          - set(body, Concat([body, " rpc_method=", attributes["rpc_method"]], "")) where IsString(body) and attributes["rpc_method"] != nil
+          - set(body, Concat([body, " expected_audience=", attributes["expected_audience"]], "")) where IsString(body) and attributes["expected_audience"] != nil
+          - set(body, Concat([body, " peer_addr=", attributes["peer_addr"]], "")) where IsString(body) and attributes["peer_addr"] != nil
+          - set(body, Concat([body, " replay_scope_bootstrap_allowed=", attributes["replay_scope_bootstrap_allowed"]], "")) where IsString(body) and attributes["replay_scope_bootstrap_allowed"] != nil
+          - set(body, Concat([body, " error=", attributes["error"]], "")) where IsString(body) and attributes["error"] != nil
+          - set(body, Concat([body, " exception_message=", attributes["exception.message"]], "")) where IsString(body) and attributes["exception.message"] != nil
 
 exporters:
   otlp/tempo: