DevOps

On-call escalation built for engineering teams

Route monitoring alerts (Grafana, Prometheus, Sentry, Uptime Kuma) to the on-call engineer through Telegram, phone calls and SMS. Acknowledgement closes the chain; resolved hook cancels it. Vacation handoff and weekly quiet hours protect responder time without dropping critical priority on the floor.

Start free Talk to engineering
API-first Webhook intake UUID identifiers

Day in the life

  1. 00:12
    Grafana
    p99 checkout latency above SLO for 5 min · firing
  2. 00:14
    WardenPoint
    Telegram voice + ringtone to primary on-call
  3. 00:18
    Engineer
    Acks from Telegram · chain cancelled · runbook opened
  4. 00:42
    Grafana
    Alert resolved · audit log writes 'resolved'

Where teams hurt

DevOps pain — and what we change

These are the three on-call complaints we hear from engineering teams switching to WardenPoint.

Alerts in Slack get missed

On-call engineers turn off Slack notifications at night. Critical alerts blend into channel noise and miss SLA.

Fix

Promote critical severity to phone + Telegram voice. Slack remains the informational channel.

PagerDuty feels heavy

Many teams adopt PagerDuty for one feature and pay for an incident-management suite they never use.

Fix

Pick only the alerting layer. WardenPoint integrates with your existing ITSM if you have one.

Escalation logic in shell scripts

Ad-hoc retry-and-page scripts grow into untested spaghetti. They break when on-call rotation changes.

Fix

Declarative escalation chains. Rotation is data, not code; tested in CI.

Escalation policy

Recipe: production on-call escalation

A policy that fits most engineering teams. Acknowledgement at any step cancels the rest.

  1. 1
    At minute
    0:00

    Telegram voice to primary

    Primary on-call gets a voice message in Telegram. Cheap, fast, bypasses DND.

    telegram_voice
  2. 2
    At minute
    0:03

    Phone call to primary

    Real PSTN call to the primary's phone. DTMF 1 acknowledges.

    voice_call
  3. 3
    At minute
    0:08

    Phone call to secondary

    Secondary on-call gets a call. Email lands on the team distribution list for situational awareness.

    voice_callemail
  4. 4
    At minute
    0:15

    Page the manager

    Engineering manager gets a phone call. The pipe is empty if no one acked within the SLA.

    voice_calltelegram_voice

Tools you bring

Wire to whatever you already run

These are the alert sources DevOps teams plug in most often.

Grafana
Webhook contact point
Prometheus
Alertmanager receiver
Sentry
Issue alerts
Uptime Kuma
Monitor webhook
Zabbix
Action media type

What changes

What you should see after switching

Wake-ups only when warranted

Critical severity hits voice; warnings stay in chat. Phone calls drop to alerts that actually need a human.

Audit-ready timeline

Every alert leaves an audit line with actor, channel, status. Post-mortems write themselves.

Less alerting surface area

Replace cron scripts, Slack hacks and partial PagerDuty configs with one declarative policy.

DevOps FAQ

Engineering questions answered

It replaces the alerting + escalation surface. Incident-management suites (postmortems, run-books, status pages) live elsewhere — we link out to yours via the audit log.
Free plan

Test your first on-call policy this afternoon

Wire one Grafana alert, set up a 4-step chain and send a fake firing. No card, no commitment.

Start free Talk to engineering
  • Free forever plan
  • Declarative policy
  • Audit log built in