Set Thresholds & Contacts¶

Configure who gets notified and under what conditions.

Alert thresholds¶

Each cabinet has four explicit temperature bands plus a target. Alerts fire whenever a reading falls outside the warning or critical band:

Band	Meaning	Default — fridge (target 4°C)	Default — freezer (target -18°C)
Warn low	Drift colder than expected	2°C	-20°C
Warn high	Drift warmer than expected	7°C	-15°C
Crit low	Critically cold	0°C	-22°C
Crit high	Critically warm	10°C	-12°C

A reading above warn_high (or below warn_low) triggers a warning. A reading above crit_high (or below crit_low) triggers a critical. This explicit-band model lets you set asymmetric ranges — useful when, for example, a display fridge can run a little colder without harm but mustn't drift warmer.

To change thresholds:

Open the cabinet detail page from the dashboard
Click Edit thresholds in the sidebar
Adjust the target and the four bands
Save

UK food safety guidance

For chilled food: keep below 8°C, target 5°C or below.
For frozen food: keep below -18°C.
The default bands follow these guidelines but you can tighten them.

Alert contacts¶

Add the people who should receive alerts:

Go to Settings → Contacts
Click + Add contact
Enter name, email address, and optionally phone number
Save

Contacts receive:

Email — all plans, sent immediately when a threshold is breached
SMS — Pro and Enterprise plans, sent after a configurable delay if the alert is unacknowledged
Voice call — Pro and Enterprise plans, escalation after SMS

Alert sensitivity¶

Each contact has an Alert sensitivity setting that controls which severity levels trigger a notification for that contact:

Setting	Behaviour
All alerts	Notified on both warning and critical alerts (default)
Critical only	Warning-level alerts are silently skipped; only critical alerts trigger a notification
None (muted)	This contact is never notified (useful for temporarily silencing without removing them)

This lets you set up a tiered contact list — for example, a general inbox on "All alerts" and a manager's phone on "Critical only" so they're only woken for genuine emergencies.

On-call schedule¶

By default a contact is available at all times. You can optionally restrict when they receive alerts by adding an on-call schedule:

Open the contact's edit modal
Expand On-call schedule (optional)
Click + Add shift and choose the day, start time, and end time
Save

Contacts with at least one shift are only notified when the current site-local time falls within one of their active windows. Cross-midnight shifts (e.g. 22:00–06:00) work as expected. A contact with no schedule entries is always available — the schedule is additive, not restrictive, so existing contacts are unaffected until you configure a schedule.

Combine schedule + sensitivity

On-call schedule and alert sensitivity work together. A contact set to "Critical only" with a Mon–Fri 09:00–17:00 schedule will only receive critical alerts during business hours. Outside those hours they won't be notified at all — make sure someone else covers the out-of-hours window.

Alert escalation¶

ChillCheck escalates alerts automatically if they go unacknowledged:

Threshold breached
      │
      │  immediately
      ▼
Email + SMS sent to all contacts                [SMS: Pro/Enterprise]
      │
      │  after call_delay_mins (default: 30 min)
      ▼
Voice call to all contacts with phone numbers   [Pro/Enterprise]

To acknowledge an alert you have three options:

Dashboard — click Acknowledge on the Alerts page and select a root cause.
Email link — every alert email contains an "Acknowledge alert" button. Clicking it acknowledges the alert immediately without logging in. The link is valid for 7 days and works once.
Voice call — press 1 during the automated call (see Voice calls section below).

Tuning the escalation timing¶

The defaults aim for a balance: email and SMS arrive immediately so you know straight away, and a voice call only follows if no-one has acknowledged after 30 minutes. You can change both delays under Settings → Alert Rules for each site.

You want	Adjust	Trade-off
Quicker escalation (e.g. you keep missing the SMS)	Lower `call_delay_mins` to 10–15 min	The phone may ring while you're still reading the SMS. Fine if your team is hands-on.
Slower escalation (e.g. SMS is enough most of the time)	Raise `call_delay_mins` to 45–60 min	A genuine emergency takes longer to escalate to a call. Only do this if your team reliably reads SMS within that window.
Email-first, SMS later (e.g. you're at a desk during the day)	Raise `sms_delay_mins` to 5–10 min	You may miss alerts when away from email. Less suitable for shop-floor teams.
Skip SMS, email + call only (rare)	Raise `sms_delay_mins` to match `call_delay_mins`	Loses a softer interruption tier. Voice call is your only mobile nudge.

Don't set the call delay shorter than 5 minutes

A very short delay means the call rings before anyone has had a chance to read the SMS and respond — you'll end up acknowledging the alert during the call instead of avoiding it. The defaults are tuned to give you a real window to react.

Acknowledging an alert¶

Clicking Acknowledge opens a modal that captures what happened. This silences further escalation and gives compliance audits the context they need.

You must choose a root cause:

Cause	When to use it
Door left open	A door was wedged or left open during stocking, delivery, or cleaning
Power outage	The cabinet lost power
Defrost cycle	A scheduled defrost ran longer or warmer than usual
Compressor fault	The compressor isn't cooling correctly
Door seal failure	A perished or damaged gasket is letting warm air in
Overstocked / airflow blocked	The cabinet was packed too tightly for cold air to circulate
Sensor placement	The sensor was knocked into a warmer spot
Sensor fault	The sensor itself is misreading (replace the battery or unit)
False alarm	The temperature was actually fine — e.g. brief spike from a hot dish being put in
Other	Anything not covered above — explain in the note field

The Notes field is optional but encouraged. Describe what action was taken, e.g. "Closed door, contents repacked, temperature back to 4°C within 15 minutes". Up to 500 characters.

Food safety outcome (HACCP corrective action)¶

Below the notes field is an optional Food Safety Outcome section. Completing it creates a formal HACCP corrective action record for the alert:

Field	Description
Was food safe?	Select Yes or No. Yes = the food was checked and found safe. No = food may have been exposed to unsafe temperatures.
Food discarded	If food was not safe, toggle this on and optionally enter the quantity (kg) and estimated cost (£).
Action taken	Free text — what was done to address the issue and prevent recurrence.
Decided by	Name or role of the person who assessed the situation.

These fields are completely optional — acknowledging an alert without completing them works the same as before. But if you do fill them in, the data appears as a badge on the alert card, is included in the alerts CSV export, and forms part of your HACCP documentation trail.

Bulk acknowledgment¶

After a connectivity outage or a defrost event that triggered several alerts at once, you can acknowledge them all in one action:

On the Alerts page, tick the checkboxes on the open alerts you want to acknowledge (or click Select all)
A bar appears at the bottom of the screen showing how many are selected
Click Acknowledge all, choose a root cause, and optionally add a note
All selected alerts are acknowledged immediately

Bulk acknowledgment applies the same root cause and note to every selected alert. It doesn't support the food safety outcome form — use individual acknowledgment for alerts where you need to record a corrective action.

Acknowledging from a voice call¶

When the call escalation fires, the spoken message ends with "Press 1 to acknowledge this alert." Pressing 1 on your handset acknowledges the alert immediately — you'll hear "Acknowledged. Thank you. Goodbye." before the call ends. Further SMS and call escalations stop, and the alert is marked acknowledged the same way as a dashboard acknowledgment.

If you don't press 1 within 10 seconds, you'll hear "No acknowledgment received. The alert remains active. Goodbye." and the call ends without acknowledging.

A phone acknowledgment doesn't capture a root cause (you can't pick a category over the phone), so the audit trail records it as "By phone" with the phone number that pressed the digit and a note saying "Acknowledged via phone +44… during voice call". If you want to fill in the root cause and a fuller note afterwards, open the alert in the dashboard.

If no-one answers (or the call goes to voicemail), the alert stays open and the next contact in priority order is called.

Muting alerts on a known-faulty cabinet¶

When a cabinet is mid-service, awaiting a part, or you already know it's running warm, you don't want every reading firing a fresh alert. Muting suppresses temperature alerts on that cabinet for a chosen period.

To mute a cabinet:

Open the cabinet detail page
Click Mute temperature alerts in the sidebar
Choose a mute-until date and time (max 90 days out)
Enter a reason — required, e.g. "Compressor service booked Friday 22nd"
Save

While muted:

High and low temperature alerts are suppressed for that cabinet
Readings are still recorded — your compliance log keeps running, the temperature chart still updates, and the dashboard card shows a Muted chip alongside the current status
Sensor offline, low battery, and weak signal alerts continue to fire — these aren't related to the cabinet itself, and you still want to know if your monitoring goes dark
Any open temperature alert stops escalating — no more SMS, no more call. The alert remains visible until you acknowledge or resolve it.

The mute lifts automatically at the chosen time, or you can unmute manually from the banner at the top of the cabinet detail page. Every mute and unmute is recorded in the audit log with the reason, who did it, and the date — so an inspector can see exactly what was happening and why.

When to mute, and when not to

Mute when you know there's a real problem and you're managing it — engineer booked, part on order, or temporarily over-stocked. The mute reason gives the audit trail context.

Don't mute as a way of silencing nuisance alerts caused by tight thresholds or poor sensor placement. Fix the underlying issue (loosen the warning band, tape the sensor in place, or follow up with the sensor-placement root cause) — muting won't make the temperature problem go away, it just hides it.

Anyone in your org — owner, admin, or member — can mute and unmute. The shop floor reality is that the person who spots the problem is often the one already there.

Commissioning grace period¶

When you assign a sensor to a cabinet for the first time, ChillCheck automatically mutes that cabinet's temperature alerts for 2 hours. Without this, every install would generate a noise alert — a fresh-out-of-the-box sensor reads ambient room temperature for an hour or two until the cabinet's compressor has had a chance to cool the air around it.

The auto-mute behaves exactly like a manual mute (you'll see the same banner on the cabinet, with reason "Commissioning grace period - first 2 hours after sensor assignment") and is recorded in the audit log. Sensor offline / battery / signal alerts are unaffected, so a broken sensor is still surfaced during commissioning.

If the cabinet was already manually muted when you assigned the sensor, the existing mute is left alone — your reason and until-time aren't overwritten. You can also unmute early at any time if the cabinet was already cold (e.g. you've reassigned an in-use sensor).

Out-of-hours behaviour¶

On Pro and Enterprise plans you can configure out-of-hours escalation:

During out-of-hours, the SMS step is skipped and escalation goes straight to a voice call
Configure your business hours in Settings → Alert Rules

Testing alerts¶

To confirm your contacts and alert delivery are working:

Go to Settings → Contacts
Click Send test email next to a contact

For a full end-to-end test, briefly warm a sensor (hold it in your hand for 5 minutes) and confirm alerts arrive as expected.

You're all set¶

Your ChillCheck installation is complete. Sensor readings are being stored, cabinets are being monitored, and your team will be alerted if anything goes wrong.

What's next: