20260721 metrics 服务新增 CCR rpki-client 格式检查指标与 Grafana 面板(#125)

This commit is contained in:
2026-07-21 18:07:05 +08:00
parent 4074323302
commit f86fbd4939
9 changed files with 956 additions and 33 deletions

View File

@@ -125,6 +125,12 @@ If a separately deployed RTR service on the same host continuously writes `rtr-s
The command accepts only an absolute path on the same host, validates the candidate Compose mount, and atomically updates `.env`. It recreates only `artifact-metrics`, Prometheus, and Grafana; it does not restart or recreate `ours-rp-soak`. The `artifact-metrics` container mounts the directory read-only and reads it through `RPKI_METRICS_RTR_REPORT_DIR`, producing `ours_rp_rtr_*` metrics. By default, `RTR_REPORT_DIR` points at an empty fallback directory under the installer data root, so the metrics container also starts cleanly without an RTR service.
## CCR artifact format check metrics
The metrics image ships the official `rpki-client 9.8`. For every newly completed run, `artifact-metrics` runs `rpki-client -f result.ccr` once and exports `ours_rp_ccr_format_check_*` metrics (cumulative pass/ccr_parse_error/tool_error counters, latest check result, check duration, consecutive failures). The Grafana `Ours RP Soak Overview` dashboard shows them in a dedicated panel row. On service start only the latest existing run is checked; older history is not backfilled.
Related environment variables (normally no need to change): `CCR_CHECK_BIN` (defaults to the in-image `/opt/ours-rp/bin/rpki-client`) and `CCR_CHECK_TIMEOUT_SECS` (default 120 seconds).
## Data Directory
Default host directory:

View File

@@ -125,6 +125,12 @@ GRAFANA_ADMIN_PASSWORD=admin
该命令只接受本机绝对路径,会先校验候选 Compose 挂载,再原子更新 `.env`。随后只重建 `artifact-metrics`、Prometheus 和 Grafana,不重启或重建 `ours-rp-soak`。`artifact-metrics` 以只读方式挂载该目录,并通过 `RPKI_METRICS_RTR_REPORT_DIR` 读取 report,输出 `ours_rp_rtr_*` 指标。默认 `RTR_REPORT_DIR` 指向安装包数据目录下的空 fallback 目录,因此未接入 RTR 服务时 metrics 容器也能稳定启动。
## CCR 产物格式检查监控
metrics 镜像内置官方 `rpki-client 9.8`。`artifact-metrics` 会对每个新完成 run 的 `result.ccr` 执行一次 `rpki-client -f` 格式检查,并输出 `ours_rp_ccr_format_check_*` 指标(累计 pass/ccr_parse_error/tool_error、最近检查结果、检查耗时、连续失败数)。Grafana `Ours RP Soak Overview` 底部有对应面板。服务启动时只补查最新的一个历史 run,不回扫全部历史。
相关环境变量(通常无需修改):`CCR_CHECK_BIN`(默认为镜像内 `/opt/ours-rp/bin/rpki-client`)、`CCR_CHECK_TIMEOUT_SECS`(默认 120 秒)。
## 数据目录
默认宿主机目录: