Why Firefly

Firefly is for Java teams that want clear scheduling semantics, a light core, and a separate operational surface.

Time Correctness

Many scheduling bugs come from default time zones, DST, deployment-machine configuration, and runtime cursors being mixed together. Firefly requires jobs to declare IANA ZoneId values and stores runtime state as UTC Instant values.

Lightweight Core

scheduler-core stays pure Java. It can be embedded in a business process or hosted by standalone Server. HTTP, JDBC, Netty, and Prometheus live outside the core.

Replaceable Modules

• integrations/* handles framework integration
• stores/* handles persistence
• transports/* handles remote executor communication
• apis/* handles management APIs
• plugins/* handles optional capabilities

Operationally Friendly

Admin UI and Admin HTTP API are separated. Java API nodes return JSON; Node Admin UI handles pages, sessions, and proxying. The Prometheus plugin independently exposes /metrics.

Not a Fit

Firefly is not a large DAG workflow engine or data orchestration platform. If your main requirements are complex dependency graphs, dataset lineage, manual approval flows, or backfill orchestration, a workflow engine may be a better fit.