feat(syslog): 优化大日志通道并增加异常连续消费兜底
- 128 KiB 及以上日志默认进入大日志 Topic - InfluxDB 原文及达梦大字段超过 256 KiB 时截断保存 - 截断内容追加 HB_XDR_TRUNCATED 标记 - InfluxDB 写入失败时继续执行规则解析和达梦写入 - 未知或系统性异常优先写入普通/大日志 DLT - DLT 写入失败时记录特殊告警并提交 offset,避免阻塞后续消息 - DLT 失败等待时间默认限制为 1 秒 - 增加连续性放行、DLT 成功及失败监控指标 - 更新 Docker Compose、v3 升级文档及相关测试
This commit is contained in:
+130
@@ -0,0 +1,130 @@
|
||||
#!/usr/bin/env bash
|
||||
set -euo pipefail
|
||||
|
||||
# Docker 单文件版:直接在 Kafka 容器内创建并校验统计投影 Topic。
|
||||
# 默认值适配测试环境;如容器名、Kafka 地址或安装目录不同,可通过环境变量覆盖。
|
||||
|
||||
DOCKER_BIN="${DOCKER_BIN:-docker}"
|
||||
KAFKA_CONTAINER="${KAFKA_CONTAINER:-kafka_ycgy-bt-kafka-1}"
|
||||
BOOTSTRAP="${BOOTSTRAP:-localhost:9092}"
|
||||
KAFKA_BIN="${KAFKA_BIN:-/opt/kafka/bin}"
|
||||
RAW_TOPIC="${RAW_TOPIC:-agent-01-syslog-topic}"
|
||||
PROJECTION_TOPIC="${PROJECTION_TOPIC:-${RAW_TOPIC}.STATS.PROJECTION}"
|
||||
RETENTION_MS="${RETENTION_MS:-604800000}"
|
||||
COMMAND_CONFIG="${COMMAND_CONFIG:-}"
|
||||
|
||||
die() {
|
||||
echo "ERROR: $*" >&2
|
||||
exit 1
|
||||
}
|
||||
|
||||
command -v "${DOCKER_BIN}" >/dev/null 2>&1 || die "未找到 Docker CLI: ${DOCKER_BIN}"
|
||||
|
||||
RUNNING="$("${DOCKER_BIN}" inspect -f '{{.State.Running}}' "${KAFKA_CONTAINER}" 2>/dev/null || true)"
|
||||
[[ "${RUNNING}" == "true" ]] || die "Kafka 容器不存在或未运行: ${KAFKA_CONTAINER}"
|
||||
|
||||
"${DOCKER_BIN}" exec "${KAFKA_CONTAINER}" test -x "${KAFKA_BIN}/kafka-topics.sh" \
|
||||
|| die "容器内找不到可执行文件: ${KAFKA_BIN}/kafka-topics.sh"
|
||||
"${DOCKER_BIN}" exec "${KAFKA_CONTAINER}" test -x "${KAFKA_BIN}/kafka-configs.sh" \
|
||||
|| die "容器内找不到可执行文件: ${KAFKA_BIN}/kafka-configs.sh"
|
||||
"${DOCKER_BIN}" exec "${KAFKA_CONTAINER}" bash -c 'exit 0' \
|
||||
|| die "容器内缺少 bash,无法执行原始升级脚本"
|
||||
|
||||
DOCKER_ENV=(
|
||||
-e "BOOTSTRAP=${BOOTSTRAP}"
|
||||
-e "KAFKA_BIN=${KAFKA_BIN}"
|
||||
-e "RAW_TOPIC=${RAW_TOPIC}"
|
||||
-e "PROJECTION_TOPIC=${PROJECTION_TOPIC}"
|
||||
-e "RETENTION_MS=${RETENTION_MS}"
|
||||
)
|
||||
KAFKA_COMMON=(--bootstrap-server "${BOOTSTRAP}")
|
||||
|
||||
if [[ -n "${COMMAND_CONFIG}" ]]; then
|
||||
"${DOCKER_BIN}" exec "${KAFKA_CONTAINER}" test -r "${COMMAND_CONFIG}" \
|
||||
|| die "容器内无法读取 COMMAND_CONFIG: ${COMMAND_CONFIG}"
|
||||
DOCKER_ENV+=(-e "COMMAND_CONFIG=${COMMAND_CONFIG}")
|
||||
KAFKA_COMMON+=(--command-config "${COMMAND_CONFIG}")
|
||||
fi
|
||||
|
||||
echo "检查 Kafka 连接及原始 topic ..."
|
||||
"${DOCKER_BIN}" exec "${KAFKA_CONTAINER}" \
|
||||
"${KAFKA_BIN}/kafka-topics.sh" \
|
||||
"${KAFKA_COMMON[@]}" \
|
||||
--describe \
|
||||
--topic "${RAW_TOPIC}" >/dev/null \
|
||||
|| die "无法连接 Kafka 或原始 topic 不存在: ${RAW_TOPIC}"
|
||||
|
||||
echo "Kafka 容器: ${KAFKA_CONTAINER}"
|
||||
echo "Bootstrap: ${BOOTSTRAP}"
|
||||
echo "原始 topic: ${RAW_TOPIC}"
|
||||
echo "投影 topic: ${PROJECTION_TOPIC}"
|
||||
echo "开始执行 Docker 版升级 ..."
|
||||
|
||||
exec "${DOCKER_BIN}" exec -i "${DOCKER_ENV[@]}" "${KAFKA_CONTAINER}" \
|
||||
bash -s <<'KAFKA_UPGRADE_SCRIPT'
|
||||
set -euo pipefail
|
||||
|
||||
if [[ "${PROJECTION_TOPIC}" == "${RAW_TOPIC}.STATS.MATERIALIZE" ]]; then
|
||||
echo "ERROR: STATS.PROJECTION 不能与 STATS.MATERIALIZE 使用同一个 Topic" >&2
|
||||
exit 1
|
||||
fi
|
||||
|
||||
TOPICS="${KAFKA_BIN}/kafka-topics.sh"
|
||||
CONFIGS="${KAFKA_BIN}/kafka-configs.sh"
|
||||
COMMON=(--bootstrap-server "${BOOTSTRAP}")
|
||||
if [[ -n "${COMMAND_CONFIG:-}" ]]; then
|
||||
[[ -r "${COMMAND_CONFIG}" ]] || { echo "ERROR: 无法读取 ${COMMAND_CONFIG}" >&2; exit 1; }
|
||||
COMMON+=(--command-config "${COMMAND_CONFIG}")
|
||||
fi
|
||||
|
||||
echo "Kafka=${BOOTSTRAP}"
|
||||
echo "原始 Topic=${RAW_TOPIC}"
|
||||
echo "统计投影 Topic=${PROJECTION_TOPIC}"
|
||||
|
||||
RAW_INFO=$("${TOPICS}" "${COMMON[@]}" --describe --topic "${RAW_TOPIC}")
|
||||
REPLICAS=$(printf '%s\n' "${RAW_INFO}" | awk '
|
||||
/ReplicationFactor:/ {for(i=1;i<=NF;i++) if($i=="ReplicationFactor:"){print $(i+1);exit}}
|
||||
')
|
||||
if [[ ! "${REPLICAS}" =~ ^[1-9][0-9]*$ ]]; then
|
||||
echo "ERROR: 无法从 ${RAW_TOPIC} 读取副本数" >&2
|
||||
exit 1
|
||||
fi
|
||||
|
||||
echo "[1/3] 创建 Topic(已经存在时不重复创建)"
|
||||
"${TOPICS}" "${COMMON[@]}" --create --if-not-exists \
|
||||
--topic "${PROJECTION_TOPIC}" \
|
||||
--partitions 1 \
|
||||
--replication-factor "${REPLICAS}" \
|
||||
--config cleanup.policy=delete \
|
||||
--config "retention.ms=${RETENTION_MS}"
|
||||
|
||||
echo "[2/3] 更新保留策略"
|
||||
if ! "${CONFIGS}" "${COMMON[@]}" --alter \
|
||||
--entity-type topics --entity-name "${PROJECTION_TOPIC}" \
|
||||
--add-config "cleanup.policy=delete,retention.ms=${RETENTION_MS}"; then
|
||||
echo "WARN: Kafka 配置更新超时;新建命令已携带相同策略,继续验证 Topic" >&2
|
||||
fi
|
||||
|
||||
echo "[3/3] 验证 Topic"
|
||||
INFO=$("${TOPICS}" "${COMMON[@]}" --describe --topic "${PROJECTION_TOPIC}")
|
||||
printf '%s\n' "${INFO}"
|
||||
PARTITIONS=$(printf '%s\n' "${INFO}" | awk '
|
||||
/PartitionCount:/ {for(i=1;i<=NF;i++) if($i=="PartitionCount:"){print $(i+1);exit}}
|
||||
')
|
||||
ACTUAL_REPLICAS=$(printf '%s\n' "${INFO}" | awk '
|
||||
/ReplicationFactor:/ {for(i=1;i<=NF;i++) if($i=="ReplicationFactor:"){print $(i+1);exit}}
|
||||
')
|
||||
if [[ ! "${PARTITIONS}" =~ ^[1-9][0-9]*$ ]]; then
|
||||
echo "ERROR: ${PROJECTION_TOPIC} 没有有效分区" >&2
|
||||
exit 1
|
||||
fi
|
||||
if [[ "${PARTITIONS}" != "1" ]]; then
|
||||
echo "WARN: Topic 已有 ${PARTITIONS} 个分区;不会修改分区数,固定 key 仍只使用其中一个分区" >&2
|
||||
fi
|
||||
if [[ "${ACTUAL_REPLICAS}" != "${REPLICAS}" ]]; then
|
||||
echo "WARN: Topic 副本数=${ACTUAL_REPLICAS},原始 Topic 副本数=${REPLICAS}" >&2
|
||||
fi
|
||||
|
||||
echo "PASS: ${PROJECTION_TOPIC} 已创建/确认。分区数=${PARTITIONS},保留时间=${RETENTION_MS}ms。"
|
||||
echo "脚本未删除消息,也未修改任何消费组 offset。"
|
||||
KAFKA_UPGRADE_SCRIPT
|
||||
Reference in New Issue
Block a user