Platform

On-call and incident response, from the first alert to the resolved incident.

The early stages decide who gets paged, by rules you can read. The later stages page them, run the incident and measure what happened.

The CallHeim loop

Ingest, reduce, decide, page, respond, close.

Select a stage to see what it does and the rule it follows.

Before the page

alerts become a decision

After the page

paging, response and closure

Learn → MTTA and MTTR from each incident’s timestamps
REDUCEstage 2

Duplicates collapse, flapping series stop paging, related alerts group.

Repeats attach to the incident that is already open, flapping alerts collapse at 4 state changes in 600 seconds, and similar titles group at a similarity of 0.6.

In the app: Noise Rules · MaintenanceSee how noise reduction works →

12 alerts in, 1 incident out

fingerprinta91f…3c
collapsed11 duplicates
attached tothe open incident

Ingest · Reduce · Decide

Before the page.

Everything between a webhook arriving and the page reaching a person.

By default, seven ordered checks

Every alert passes through these checks, in order, before it can open an incident. Each rule is shown with its default. Orchestration rules run first, and your own noise rules can add further steps.

The decision pipeline. By default seven ordered checks run on every alert; orchestration rules run before them and your own noise rules can add more. Most checks divert the alert out of the sequence — a flapping series with nothing open to join is the one exception that still creates an incident.

  1. Maintenance window

    Suppress if a window covering this service, team or the whole organization is open right now.

    active windowOutcome: suppress

  2. Suppress rules

    Suppress if one of your own noise rules matches the alert.

    rule matchOutcome: suppress

  3. Fingerprint dedup

    Attach a repeat to the open incident in the same service. When the sender supplied a dedup key, that holds at any age. When the identity is only the title or e-mail subject, it holds for 24 hours. A P1 or P2 repeat on an open P3 to P5 incident opens a new incident instead.

    open incident; 24 h if title-keyedOutcome: attach

  4. Flap collapse

    A series that keeps changing state attaches to the open incident instead of paging again. With nothing open to join, it still opens an incident.

    4 changes / 600s (fixed)Outcome: group

  5. Title correlation

    Group with an open, still-Triggered alert from the same source and the same service when their titles overlap enough.

    similarity ≥ 0.6, 600sOutcome: group

  6. Group rules

    Apply your own grouping rules over the window.

    600s windowOutcome: group

  7. Create incident

    Nothing above matched: open the incident with the source’s severity (or the severity an orchestration rule sets), pick the escalation policy for the service, and page.

    elseOutcome: create

Defaults shown above. Orchestration rules run first and your own noise rules can add further steps.

See how noise reduction works →Browse the alert sources CallHeim accepts →

Walkthrough

Watch an alert become an incident.

One alert, walked through the checks above with their default thresholds.

01 · A webhook arrives

Inbound webhookdatadog
{
  "alert_type": "error",
  "title": "checkout-api p99 latency > 2s",
  "host": "checkout-api-canary",
}

02 · The alert gets a fingerprint

Decision receipt
fingerprint
organization · source · dedup key

The alert’s dedup key or identity, falling back to its title. The service is not an input.

dedup
open incident

A repeat joins the incident that is still open for it: at any age when the sender supplies its own dedup key, within 24 hours when the identity is only the title or e-mail subject.

03 · Duplicates collapse, grouping does not cross sources

Decision receipt
repeats
same fingerprint

Attach to the open incident with no new page. A P1 or P2 repeat on an open P3 to P5 incident opens a new incident and pages.

title grouping
title similarity ≥ 0.6

Same source, same service, still Triggered, within 600s. Never across sources.

04 · Severity is the source's, plus a suggestion

Decision receipt
incident severity
P1

Mapped from what Datadog sent (or an orchestration override).

suggestion
P1, matched “latency”, “p99”

A separate suggestion from a keyword ruleset and your resolved incidents. A person applies it; it never changes severity on its own.

05 · Routed and paged

ExampleSeverity P1Status Triggered#2041

checkout-api p99 latency > 2s

service checkout-apipolicy checkout-api — default
The service’s escalation policy pages tier 1. If nobody answers, the page moves to the next tier and ends at the mandatory fallback.

How incident severity levels work →What an escalation policy is →How severity suggestions are calculated →

Page

Escalation that keeps going until someone answers.

An unanswered tier times out and the escalation moves to the next tier, ending at the fallback.

  • Escalation policies have up to 8 tiers, each paging a person, a team or a schedule, with per-tier timeouts from 1 to 240 minutes, optional repeats and a mandatory fallback. The whole chain must fit within 24 hours.
  • Acknowledging an incident stops the escalation.
  • Each alert source is bound to a service, and the service’s escalation policy (or its team’s) sets who is paged.

See alerting and escalation policies →How pages reach your team →See on-call scheduling →

Example escalation policy
T+0:00

Tier 1

The responder on call right now, paged on their own notification order.

  1. T+0:00

    Tier 1

  2. T+5:00

    No acknowledgment → tier 2

  3. T+10:00

    Still nothing → the tier chain repeats

  4. Repeats exhausted

    One fallback target, once

  5. Acknowledged

    The walk stops

Pricing

Included on every plan.

Alerting, on-call, escalation, noise reduction, incident response, public status pages and core analytics are on every plan.

A few features depend on plan, including the REST incident API, related-incident insights, audit-log export and email-to-alert.

See what each plan includes →Read about security and trust →

CallHeim

Put a rule you can read between your alerts and your on-call.

See what each plan includes.