A collection of +180 reusable WebAssembly modules for building composable CI/CD pipelines with FluentCI fluentci.io/modules
README.md
# Dragonfly service Run a background Dragonfly test service through FluentCI's native workspace service manager and Docker backend. Typed PureScript Wasm component; requires FluentCI **v0.17.4+**, with the native services apply/status lifecycle API. ```sh fluentci run dragonfly start --version v2.0.2 fluentci run dragonfly status fluentci run dragonfly logs --tail 50 fluentci run dragonfly logs --tail 50 --follow true fluentci run dragonfly restart fluentci run dragonfly stop ``` For live logs, use `logs --follow true`. Press Ctrl-C to stop following; the background service keeps running. The native shorthand is `fluentci logs module-dragonfly-default -f` in the same project workspace. ## Default profile - Image: `ghcr.io/dragonflydb/dragonfly:v2.0.2`. - Primary port: `6379`; exposed ports: `6379`. - Readiness: tcp /; deadline 60000 ms. Single-node Redis-compatible test cache without authentication. Uses epoll instead of io_uring for standard Docker sandbox compatibility, one I/O thread and a 256 MiB memory limit. Readiness checks the allocated loopback binding; full E2E also requires a Redis PONG. Mount /data for snapshots; explicit named or bind-mounted data survives stop. These settings target disposable CI workloads. [Upstream configuration](https://github.com/dragonflydb/dragonfly/blob/v2.0.2/README.md). ## Typed configuration and lifecycle --version selects the image tag. --image accepts a complete image/digest reference; do not combine them. Start waits for readiness and returns name, state, backend, host, primary port, endpoint and every container/host port binding. Dynamic ports are returned as allocated; do not assume the host port equals the container port. Restart uses the saved native definition. Reconfiguration requires stop first. Repeated unchanged start reuses the service, and repeated stop is safe. Stop also affects native dependents, while explicit bind/named volume data is preserved. Use project to select the same workspace across calls. Docker must be installed and its daemon and lo…