Point a Datadog monitor webhook at OpsPing and every trigger pages your on-call: the monitor's title becomes the alert message, its body the description, and its alert ID a stable alias — so a Recovered transition closes the alert the Triggered one opened, automatically.
Field mapping
| Datadog field | OpsPing alert |
|---|---|
title | Message |
body | Description |
alert_id | Alias (datadog-<id>) — stable across transitions. Multi-alert monitors (grouped, e.g. by {host}) share one $ALERT_ID, so alert_scope is hashed into the alias (datadog-<id>-<hash>) and each group gets its own alert; a group's Recovered closes only that group |
alert_type | Priority: error → P2, warning → P3, info and success → P4, anything else (or missing) → P3. Datadog's own $ALERT_PRIORITY is not read, and no Datadog value maps to P1 |
Recovered transition | Closes the alert opened by Triggered |
tags | Alert tags (comma-separated; first 20 kept, each capped at 50 characters) |
event_type, host, alert_id, alert_scope, alert_type, alert_transition | Alert details |
Setup
- In OpsPing, open Integrations and create an inbound integration with the Datadog source type, pick a default service, and copy its webhook URL (
https://api.opsping.com/v2/webhooks/inbound/<id>/<token>). Treat the URL as a secret: the token in it is the credential. On save, OpsPing generates and displays an OpsPing API key (pager-…) bound to the integration — that is the key to use if you also call the Opsgenie-compatible API. Nothing you or OpsPing stores is ever sent to Datadog; webhook deliveries authenticate with the token in the URL. (Older versions of the form showed an "API key" input; that value was stored but never used.) (The OpsPing API Key source type works too; the payload below is auto-detected as Datadog either way.) - In Datadog, go to Integrations → Webhooks, add a new webhook named
opsping, and set its URL to the OpsPing webhook URL. - Replace the webhook's Payload with the JSON below. Datadog's default payload contains no
alert_idoralert_transition, so without this step alerts would have no alias: no deduplication and no automatic close. The variable names are Datadog's (Datadog webhook variables ↗); the JSON keys are what OpsPing reads. Note there is no$ALERT_BODYvariable — the event text is$EVENT_MSG.
{
"title": "$EVENT_TITLE",
"body": "$EVENT_MSG",
"alert_id": "$ALERT_ID",
"alert_scope": "$ALERT_SCOPE",
"alert_transition": "$ALERT_TRANSITION",
"alert_type": "$ALERT_TYPE",
"event_type": "$EVENT_TYPE",
"host": "$HOSTNAME",
"tags": "$TAGS"
}- Mention
@webhook-opspingin the message of every monitor that should page. ARecoveredtransition on the same monitor closes the alert itsTriggeredtransition opened.
Operational notes
- Adding
alert_scopeto an existing payload changes the alias: alerts still open under the plaindatadog-<id>alias will not be closed by aRecoveredcarrying the new scoped alias. Close those open alerts by hand in OpsPing once, right after switching the payload. - Keep Datadog on its own OpsPing integration (one integration per source system) so provenance, rate limits, and team scoping stay clean.
- Route noise before it pages: alert policies can suppress, delay, or auto-ack Datadog-sourced alerts by source, tags, or priority — with a whole-chain dry-run first.
- Tools that can only send mail get an email-to-alert address instead, with
[P1]–[P5]subject tags.
Why a webhook and not Datadog’s Opsgenie tile?
Datadog’s Opsgenie integration only lets you choose Opsgenie’s US or Europe region (Datadog docs); it has no custom API URL, so it can’t be pointed at OpsPing. Use the Webhooks integration above. When you cut over, replace @opsgenie-<service> handles in monitor messages with the matching @webhook-<name> handle.
FAQ
Does the Datadog mapper handle monitor recovery?
Yes — aliasing is what closes the loop: alert_id maps to a stable alias, and a Recovered transition closes the alert that the earlier Triggered opened, even if it was already acknowledged.
Do repeat Datadog notifications create duplicate alerts?
No. If an alert with the same alias is already active, a repeat firing updates that alert and increments its occurrence count instead of creating a second row. It can send another push unless the alert is acked, snoozed, or covered by a maintenance window.
Where do I change the Datadog webhook if the token leaks?
Rotate it from the integration's page in OpsPing — the old URL stops working immediately. The URL contains the token, so treat it like a password.