Healthcare — pharmacy cold chain

The problem. Vaccines, blood products and many medications are only safe within a narrow temperature band. A pharmacy or ward fridge that drifts out of range — a door left ajar, a failing compressor overnight — can spoil tens of thousands of euros of stock and, worse, leave a patient given a product that quietly lost its potency. Staff can't watch every fridge; they need an alarm the moment one strays, and a clean temperature record to prove the chain held.

What we solve here is protecting temperature-sensitive stock with live monitoring and an auditable record.

Set up demo data

New workspace? Run this once (Python) to create a fridge's temperature series, create the subscription before we listen, and push it above the 2–8 °C band. Safe to re-run.

import datahub_sdk, numpy as np, pandas as pd

client = datahub_sdk.DataHubClient.from_env()

client.timeseries.create([datahub_sdk.TimeSeries(external_id="fridge_pharmacy_12_temp_c", name="Pharmacy fridge 12 temp", unit="deg_c", value_type="float")])
client.subscriptions.create([datahub_sdk.Subscription(
    external_id="pharmacy_fridges", name="Pharmacy fridges", timeseries=["fridge_pharmacy_12_temp_c"])])
idx = pd.date_range(end=pd.Timestamp.now(tz="UTC"), periods=60, freq="1min")
t = np.full(60, 4.5); t[-10:] = 9.3                 # rises above the 8 °C ceiling
client.timeseries.insert_from_lists(timestamps=idx, values=t, ts="fridge_pharmacy_12_temp_c")

1. Watch every fridge live

Each fridge streams its temperature; a subscription brings them into one feed. When one leaves its band, raise a cold_chain_excursion event so it's attended before stock is lost. See Consume live data.

Java

import ai.intellistream.datahub.sdk.subscriptions.SubscriptionMessage;

try (var stream = client.subscriptions().listen(List.of("pharmacy_fridges"))
        .stream((SubscriptionMessage msg) -> {   // auto-acks after each message
            if (outOfBand(msg.payload(), 2.0, 8.0)) {     // °C — vaccine fridge band
                EventModel excursion = new EventModel();
                excursion.setExternalId("cold_chain_excursion_f12_" + System.currentTimeMillis());
                excursion.setType("cold_chain_excursion");
                excursion.setStatus("critical");
                excursion.setMetadata(Map.of("fridge", "fridge_pharmacy_12", "temp_c", "9.3"));
                excursion.setEventTime(ZonedDateTime.now());
                client.events().create(List.of(excursion));
            }
        })) {
    awaitShutdown();   // your app lifecycle; closing the stream ends delivery
}

Python

with client.subscriptions.listen(["pharmacy_fridges"]) as listener:
    for msg in listener:
        if out_of_band(msg.payload, 2.0, 8.0):         # °C — vaccine fridge band
            client.events.create([datahub_sdk.Event(
                external_id=f"cold_chain_excursion_f12_{int(pd.Timestamp.now().timestamp())}",
                type="cold_chain_excursion", status="critical",
                event_time=pd.Timestamp.now(tz="UTC"),
                metadata={"fridge": "fridge_pharmacy_12", "temp_c": "9.3"})])
        listener.ack([msg.message_id])

Rust

use dataplatform_rust_sdk::events::Event;
use chrono::Utc;

let mut listener = api.subscriptions.listen(&["pharmacy_fridges"]).await?;
while let Some(Ok(msg)) = listener.next().await {
    if out_of_band(&msg.payload, 2.0, 8.0) {           // °C — vaccine fridge band
        let mut excursion = Event::new(format!("cold_chain_excursion_f12_{}", Utc::now().timestamp()));
        excursion.r#type = Some("cold_chain_excursion".into());
        excursion.status = Some("critical".into());
        excursion.add_metadata("fridge".into(), "fridge_pharmacy_12".into());
        excursion.add_metadata("temp_c".into(), "9.3".into());
        excursion.set_event_time(Utc::now());
        api.events.create(&vec![excursion]).await?;
    }
    listener.ack(&[msg.message_id.as_str()]).await?;
}

2. Keep the compliance record

For an audit, retrieve a fridge's temperature history over any window to produce the time-stamped record showing the chain held. See Query & aggregate.

See the result

The warming fridge trips the loop while there's still time to act:

cold_chain_excursion_f12_…  → critical   (9.3 °C, above the 2–8 °C band)

See also

• Consume live data — the fridge-monitoring loop.
• Turn readings into events — the excursion rule.
• Maritime — reefer cold chain — the same pattern at sea.