On-call without PagerDuty: lightweight escalation for small teams

A single Slack channel that pings everyone is not on-call — it is hope. When something breaks at 2 a.m., "@here" reaches five people who all assume one of the other four has it, and the one person who could actually fix it has notifications muted. Real on-call needs a chain: alert someone, and if they do not respond, alert the next person, and keep going until a human actually acknowledges it. The good news is you can get proper on-call escalation without PagerDuty and its per-seat pricing. A small team needs three things — tiers, delays, and acknowledgement — and none of them require a heavyweight platform.

What "escalation" actually means

Escalation is just a fallback sequence with timers. The shape is always the same:

• Tier 1: notify the primary on-call person.
• Wait: give them a few minutes to acknowledge.
• Tier 2: no ack? Notify the backup (or the whole team).
• Repeat: keep nudging until someone says "I've got this."

That loop is the entire feature. Everything else PagerDuty sells is polish on top of it — and polish you may not need yet.

The three pieces a small team really needs

You can skip 90% of an enterprise on-call product if you have these:

1. Tiers. A defined order of who gets told, not a free-for-all.
2. Delays. A grace period between tiers so a quick responder is not immediately joined by the whole company.
3. Acknowledgement. A way for someone to claim an incident and stop the escalation, so a handled problem goes quiet instead of paging the backup anyway.

Miss acknowledgement and you get the worst outcome: people learn the alerts are noise and start ignoring all of them.

Doing it without a heavy platform

Lookout includes native escalation policies built for exactly this. You define tiers with delays between them, and an incident climbs the chain until it is acknowledged. Acknowledgement works from a signed link in the alert itself — tap it from your phone and the escalation stops, no login dance at 2 a.m. If a tier goes unanswered, it repeats and moves up. You get the PagerDuty loop without the PagerDuty bill.

And if you do already live in PagerDuty or Opsgenie, Lookout integrates with both — severity-mapped and deduplicated — so you can let those own escalation and use Lookout to decide how bad something is. Use the native policies when you do not want another vendor; hand off to the dedicated platform when you already pay for one.

Make alerts worth acknowledging

Escalation only works if the alerts deserve a response. Two rules keep the chain credible:

• Threshold, don't trickle. Fire on a spike over a 1h or 24h window, not on every single event, so the page means "this is real."
• Route by severity. A fatal crash or an SLO burn should escalate; a new low-priority issue can sit in Email or Slack and wait for morning.

Get this right and people trust the page enough to actually wake up for it.

Start small

You do not need a full rotation on day one. Start with two tiers — you, then the team — a five-minute delay, and acknowledgement on. That alone beats a silent channel. Add a real rotation when there are enough humans to share the nights.

The takeaway

On-call escalation without PagerDuty is well within reach: tiers, delays, and acknowledgement are the whole game, and you can run them natively or hand off to a platform you already pay for. The goal is simple — the right person wakes up, claims it, and everyone else stays asleep. Set up Lookout free — 10,000 events a month, no credit card — and give your alerts a chain that actually ends with a human.