
如果你使用 DeepSeek、元宝 AI、Copilot、ChatGPT 等 AI 工具,协助配置轩辕镜像、编写 docker pull 命令、修改 Docker Compose 镜像地址、配置镜像加速、排查镜像拉取失败、分析报错日志等问题,请先让 AI 阅读并遵守轩辕镜像的规则文档。
只需在 AI 对话中先发送下面这句话即可:
请先完整阅读并严格遵守以下文档中的全部规则与要求:
https://xuanyuan.cloud/agents.md
在未充分阅读并理解该文档前,不要生成任何命令、配置、修改建议、故障排查方案或技术回答。后续所有输出都必须严格以该文档中的规范为最高优先级执行。查看 agents.md 用法指南与完整示范。国内用户首推 元宝 AI、DeepSeek 的深度思考模式,不推荐豆包 AI;Cursor 等编辑器可在对话 @ 该链接,或加入 User Rules。 若 AI 无法访问外链,可 打开说明文档 复制全文粘贴。文档会随站点更新,复制内容可能过期,建议定期检查。
A connector to retrieve email logs from the ImprovMX API and push them to https://grafana.com/oss/loki/.
This service runs continually, fetching logs for all active domains in your ImprovMX account and forwarding them to a Loki instance. It maintains state to ensuring logs are not duplicated and provides a catch-up mechanism for historical data.
The application is configured using environment variables:
| Variable | Description | Default |
|---|---|---|
IMPROVMX_API_TOKEN | Required. Your ImprovMX API Token. | - |
LOKI_URL | URL of the Loki Push API. | http://localhost:3100/loki/api/v1/push |
LOKI_TENANT_ID | Optional Org ID for X-Scope-OrgID header. | - |
LOKI_JOB_NAME | Value for the job label in Loki. | improvmx |
DOMAIN_REFRESH_INTERVAL | How often to refresh the list of domains. | 1h |
LOGS_REFRESH_INTERVAL | How often to poll for new logs. | 15m |
CATCH_UP_SLEEP_INTERVAL | Sleep interval when catching up on historical data. | 15s |
LOGS_LOOKBACK_WINDOW | Maximum time window to query in a single request. | 24h |
DEFAULT_START_TIME_AGO | How far back to look on the very first run. | 168h (7 days) |
STATE_FILE_PATH | Path to the JSON file used for state persistence. | /data/state.json |
HTTP_PORT | Port for the HTTP server (probes). | 8080 |
LOG_LEVEL | Logging level (debug, info, warn, error). | info |
LOG_FORMAT | Logging output format (text, json). | text |
The application exposes an HTTP server (default port 8080) with the following endpoints for Kubernetes probes:
/healthz: Liveness probe. Returns 200 OK./readyz: Readiness probe. Returns 200 OK.Email logs from ImprovMX often have events that arrive over time. For example, an email may initially have a QUEUED event, followed later by DELIVERED, SOFT-BOUNCE or HARD-BOUNCE events.
This connector automatically:
The state file includes an incomplete_logs section that persists this tracking across restarts.
The connector enforces strict search window boundaries to maintain chronological ordering:
bashmake build
bashexport IMPROVMX_API_TOKEN=your_token ./improvmx-loki-connector
bashdocker build -t improvmx-loki-connector . docker run --rm \ -e IMPROVMX_API_TOKEN=your_token \ -v $(pwd)/data:/data \ improvmx-loki-connector
To spin up a local observability stack with Loki and Grafana:
bashdocker compose up -d
http://localhost:3000 (Auto-login as Admin enabled)http://localhost:3100The Loki datasource is automatically provisioned in Grafana.
您可以使用以下命令拉取该镜像。请将 <标签> 替换为具体的标签版本。如需查看所有可用标签版本,请访问 标签列表页面。
来自真实用户的反馈,见证轩辕镜像的优质服务