reddit · SideProject · 11-12 03:50 58 分 ← 返回列表

I built a tool that tells you which deploy made a Postgres query slower

When a query gets slow, our postmortems always start the same way: someone opens Grafana, someone else scrolls through Argo CD history, and we try to match timestamps by eye. I got tired of that and wrote something to do it automatically. It reads pg_stat_statements, picks up deploy events from Argo CD, Argo Rollouts, Flux or a webhook, and checks whether a latency shift in any query lines up with a rollout. If it does, you get the query, the deploy, before/after latency and how confident i

五维评分

付费意愿

需求真实度

个人可落地

竞争空间

新鲜度

AI 分析

痛点：线上 Postgres 查询变慢时，排查需要人工比对 Grafana 时间线和 Argo CD 部署历史，定位哪次发布导致了性能回归极其耗时

目标用户：SRE、平台工程/DevOps 团队、使用 Kubernetes + Argo CD/Flux 且以 Postgres 为核心数据库的中小型 SaaS 公司

证据：When a query gets slow, our postmortems always start the same way: someone opens Grafana, someone else scrolls through Argo CD history, and we try to match timestamps by eye. I got tired of that ... It's a side project and I'm looking for people willing to try it on a staging cluster.

变现：开源核心 + 托管 SaaS 订阅（按监控的 Postgres 实例数/查询数计费）；或提供团队版高级功能（多集群、历史回归知识库、自动 canary 回滚、SSO），定价 29-99 美元/月/实例

MVP：两周版：只做单一接入路径——读取 pg_stat_statements 的按 queryid 聚合延迟，接收一个通用 webhook 部署事件，做时间窗口内的前后延迟对比与简单显著性判断，命中后发一条 Slack 消息（含 query 文本、部署 ID、before/after p95、置信度）。不做 CRD、不做自动回滚，用 Python/Go 单二进制 + SQLite 存储即可。

竞品：Datadog APM + Deployment Tracking、Grafana Cloud / Grafana annotations、pganalyze、SolarWinds Database Performance Analyzer、New Relic change tracking、pgwatch / pg_stat_statements 自建看板

摘要：开发者发布了一个副业项目 pg-regression-radar，它读取 pg_stat_statements 并从 Argo CD、Argo Rollouts、Flux 或 webhook 获取部署事件，自动判断某条 Postgres 查询的延迟劣化是否与某次发布时间吻合，结果推送到 Slack/Teams/PagerDuty，用 Argo Rollouts 时还可中止 canary，可单二进制或 CRD 控制器运行。痛点真实且是 B2B 场景，但帖子只是求 staging 环境试用反馈，没有任何付费意愿原话或收入数据，且 Datadog、Grafana、pganalyze 等已覆盖部分能力，需靠 Postgres 垂直深度和轻量部署取胜。