// the find
zhp8341/flink-streaming-platform-web
基于flink的实时流计算web平台
A web UI for managing Apache Flink streaming jobs via SQL, letting you configure, start/stop, and monitor jobs through a browser instead of writing Java code. Aimed at data engineers who know Flink SQL but want to avoid the operational friction of raw Flink deployment. Pinned to Flink 1.16.x and primarily documented in Chinese.
Savepoint management (auto and manual backup + restore) is a feature most internal Flink tooling skips entirely. Supports all four Flink deployment modes (local, yarn-per-job, standalone, application), which covers most real production setups. The JAR management system letting you host connectors over HTTP per-job rather than dumping everything into flink/lib is genuinely useful for avoiding classpath conflicts. DingTalk alerting with auto-restart is a real operational need in Chinese data teams.
Flink 1.16.7 is two major versions behind (current is 1.20); the README still references 1.14/1.13 connector URLs, suggesting the docs haven't kept pace with the code. SELECT queries are explicitly not supported — you can only run INSERT pipelines, which makes ad-hoc debugging impossible from the UI. The project README is almost entirely Chinese and links to paid support for anything non-trivial, which signals limited community investment in English-language adoption. No RBAC or multi-tenancy — it's a single-user admin panel, so running it in a shared team environment requires you to build access controls yourself.