# VDIFPacker — Services

<!-- AI-GENERATED:START name="services" -->
The pipeline runs as a single Docker container managed by the EDD Master Controller.

## Container

| Property | Value |
|---|---|
| Container name | `vdif_packer` |
| Docker image | `edd_dbbc` |
| Start command | `edd_vdif_packer --host=<edd_subnet> --port=<bind_port>` |
| Registered with Master Controller | Yes (`register_as_pipeline: True`) |
| Restart policy | Controlled by Master Controller |

## Sidecar Services

| Sidecar | Purpose |
|---|---|
| Redis | Pipeline coordination and data store |
| InfluxDB | Time-series metrics |

## Grafana Dashboards

| Panel | Contents |
|---|---|
| `pipeline_status.json` | KATCP state machine state, error indicators |
| `mkrecv_status.json` | SPEAD2 capture statistics |
| `dada_monitor.json` | PSRDADA ring buffer fill levels |
| `visualizations.json` | Per-channel signal level plots |
| `waterfall.json` | Spectrogram / waterfall display |

Dashboard variables are loaded from `visualization_vars.json` and `waterfall_vars.json`.
The dashboard refresh interval is 30 s.

## Ports

| Port | Protocol | Direction | Description |
|---|---|---|---|
| `bind_port` (default 0) | TCP | inbound | KATCP control |
| configured via `input_data_streams` | UDP multicast | inbound | SPEAD2 input |
| `udp.dest_port` (default 4991) | UDP | outbound | VDIF frames to jive5ab |
<!-- AI-GENERATED:END -->