Last updated
OpsPing accepts a documented subset of the Opsgenie Alert API v2 at https://api.opsping.com: create alerts with the Opsgenie JSON shape and alias deduplication, fetch an alert and run the common actions by alert ID, tiny ID, or alias (identifierType=alias), and find alerts with alias: and status: list-query tokens. Authenticate with a new OpsPing API key using the familiar GenieKey header. Most senders change only the URL and key, and Alertmanager and Grafana can keep auto-resolve on, because they close alerts by alias. Unacknowledge, unassign, reopen, unsnooze, mark-read and delete still take the alert ID or tiny ID.
Base URL and authentication
| Opsgenie | OpsPing | |
|---|---|---|
| Base URL | https://api.opsgenie.com or https://api.eu.opsgenie.com | https://api.opsping.com (or your private deployment's hostname) |
| Header | Authorization: GenieKey <key> | Authorization: GenieKey <key>, Authorization: OpsPingKey <key>, or X-OpsPing-Key: <key> |
| Key | Opsgenie integration or API key | A new OpsPing API key — POST requests need write scope |
| Create response | 202 with a requestId to poll | 202 with the alertId included |
Supported endpoints
“Supported” means the path, method and core request shape match Opsgenie's Alert API v2. Response bodies are OpsPing's and are not guaranteed to be field-for-field identical to Opsgenie's.
| Endpoint | Status | Notes |
|---|---|---|
POST /v2/alerts | Supported | Opsgenie-shaped body; deduplicates on alias |
GET /v2/alerts | OpsPing filters | Understands alias: and status:; other text is substring search |
GET /v2/alerts/{id} | Supported | By alert ID, tiny ID, or alias with ?identifierType=alias |
GET /v2/alerts/count and /counts | Supported | /count is unfiltered; /counts accepts alias/status filters |
| Common alert actions | Supported subset | Acknowledge, close, snooze, notes, tags, assign, responders, escalate, details and priority accept IDs/tiny IDs; alias works on supported actions |
PUT /v2/alerts/{id}/priority | Supported | Accepts P1–P5 by ID, tiny ID or alias; does not re-run policies or escalation |
POST /v2/alerts/{id}/responders | Supported | Accepts responder(s); team, schedule and escalation references resolve by tenant-scoped ID or unambiguous name, subject to team access |
| Unacknowledge, unassign, reopen, unsnooze, mark-read, delete | Partial | Use alert ID or numeric tiny ID, not alias |
GET /v2/alerts/requests/{requestId} | Not supported | Creation is synchronous; use returned alertId |
GET /v2/heartbeats/{name}/ping, POST, PUT, PATCH | Supported | All four ping methods use write authorization; PATCH /{name} still updates configuration |
/v1/incidents and configuration APIs | Not compatible | Incidents, schedules, teams and policies use OpsPing-native schemas |
OpsPing also adds POST /v2/alerts/{id}/reopen and /unsnooze, which Opsgenie does not have. Every endpoint is documented in the API reference.
Fields accepted on POST /v2/alerts
| Field | Rules in OpsPing |
|---|---|
message | Required. 1–130 characters |
alias | Up to 512 characters. Deduplication key: a new create with the alias of an open or acknowledged alert increments that alert's count instead of paging again |
description | Up to 15,000 characters |
priority | P1–P5; defaults to P3 |
source | Up to 100 characters |
entity | Up to 512 characters |
tags | Up to 20 strings, each up to 50 characters |
details | Object of string values |
responders | Up to 50. user by id or username (email); team, schedule, escalation by ID or unambiguous tenant-scoped name, subject to team access |
visibleTo | user or team entries with an id |
actions | Up to 10 strings, each up to 50 characters |
user | Up to 100 characters; recorded as the actor |
note | Up to 25,000 characters |
owner | Must reference an OpsPing user in your tenant |
OpsPing-only optional fields: serviceId (link the alert to an OpsPing service), runbookUrl, recurrenceKey, and skipEscalation. A request that breaks a limit — for example, 21 tags — is rejected with 400 and a message naming the field, rather than being truncated.
The “supported subset” caveat
- Alias covers lookups and the common actions, not every action.
GET /v2/alerts/{alias}?identifierType=aliasreturns the open alert for that alias, otherwise the newest closed one, and the list query (andGET /v2/alerts/counts) acceptsalias:<value>andstatus:<value>tokens. With?identifierType=aliasyou can acknowledge, close, add a note, add or remove tags, assign, add a responder, escalate, snooze and add details. Unacknowledge, unassign, reopen, unsnooze, mark-read and delete still take thealertIdreturned on create or the numeric tiny ID. - No request polling.
GET /v2/alerts/requests/{requestId}is not implemented; creation is processed synchronously. Priority updates accept P1–P5 by ID, tiny ID or alias and do not re-run policies or escalation. - Responders must exist in OpsPing. Teams, schedules and escalations resolve by ID or unambiguous tenant-scoped name, subject to team access; users resolve by ID or email username. Missing or ambiguous names are rejected.
- Action bodies take OpsPing references. Assign accepts Opsgenie's
ownerobject (idorusername), adding responders acceptsresponderorresponders, and tag removal accepts?tags=,?tag=or a JSON body. Escalate advances the alert's own escalation policy and acceptsescalation.idonly for that current policy. - List queries are simpler. Beyond
alias:andstatus:tokens, the Opsgenie query language is not implemented: other text is a substring search, and unsupported query parameters return400. - Configuration APIs differ. Schedules, teams, escalations and policies exist in OpsPing's API with their own schemas. Use the importer rather than replaying Opsgenie configuration calls.
Test every integration before cutover: send a firing and a resolution and confirm the result in OpsPing.
Examples
Prometheus Alertmanager (opsgenie_configs)
Change api_url and api_key. Keep the trailing slash: Alertmanager appends v2/alerts to api_url. Remove any responders that name Opsgenie teams by name, or replace them with OpsPing IDs.
# alertmanager.yml — before (Opsgenie)
receivers:
- name: opsgenie
opsgenie_configs:
- api_url: https://api.opsgenie.com/
api_key: <opsgenie-api-key>
# alertmanager.yml — after (OpsPing): only the URL and key change
receivers:
- name: opsgenie
opsgenie_configs:
- api_url: https://api.opsping.com/
api_key: <opsping-api-key>
# Resolved notifications close the alert by alias
send_resolved: true
Alerts are created and deduplicated per alert group, and with send_resolved: true (Alertmanager's default for opsgenie_configs) a resolution closes the matching OpsPing alert by alias. If you prefer one OpsPing alert per Prometheus alert, keyed by fingerprint, use the native mapper instead:
# Alternative: OpsPing's native Alertmanager mapper
receivers:
- name: opsping
webhook_configs:
- url: https://api.opsping.com/v2/webhooks/inbound/<integration-id>/<token>
send_resolved: true
Full field mapping: Alertmanager integration.
Grafana (Opsgenie contact point)
In Alerting → Contact points, edit the Opsgenie contact point: set Alert API URL to https://api.opsping.com/v2/alerts, paste a new OpsPing key as the API key, and leave Auto close incidents on so resolved alerts close by alias. Or provision it from a file:
# Grafana contact point provisioning (provisioning/alerting/*.yaml)
apiVersion: 1
contactPoints:
- orgId: 1
name: opsping-opsgenie-compatible
receivers:
- uid: opsping_genie
type: opsgenie
settings:
apiUrl: https://api.opsping.com/v2/alerts # was https://api.opsgenie.com/v2/alerts
apiKey: <opsping-api-key> # new OpsPing key, not the Opsgenie key
autoClose: true # resolved alerts close by alias
sendTagsAs: details # keeps label sets within the 20-tag limit
Grafana can send every alert label as a tag; with sendTagsAs: tags or both, an alert with more than 20 labels, or a key:value label longer than 50 characters, is rejected by the tag limits, so details is the safer setting. To map each Grafana alert by its fingerprint instead, use a Grafana webhook contact point with OpsPing's Grafana integration.
Scripts and custom tools (curl)
curl -X POST https://api.opsping.com/v2/alerts \
-H "Authorization: GenieKey $OPSPING_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"message": "Disk usage above 90% on db-1",
"alias": "db-1-disk-usage",
"description": "/var/lib/postgresql at 91%",
"priority": "P2",
"source": "cron-check",
"entity": "db-1",
"tags": ["database", "disk"],
"details": {"mount": "/var/lib/postgresql", "threshold": "90"}
}'
# 202 Accepted — OpsPing returns the alert ID directly:
# {"result": "...", "alertId": "<id>", "requestId": "...", "took": 0}
# Acknowledge, add a note, then close — by alert ID (or numeric tiny ID)
curl -X POST https://api.opsping.com/v2/alerts/$ALERT_ID/acknowledge \
-H "Authorization: GenieKey $OPSPING_API_KEY" \
-H "Content-Type: application/json" -d '{"user": "deploy-bot"}'
curl -X POST https://api.opsping.com/v2/alerts/$ALERT_ID/notes \
-H "Authorization: GenieKey $OPSPING_API_KEY" \
-H "Content-Type: application/json" -d '{"note": "Rotated logs, usage at 70%"}'
curl -X POST https://api.opsping.com/v2/alerts/$ALERT_ID/close \
-H "Authorization: GenieKey $OPSPING_API_KEY" \
-H "Content-Type: application/json" -d '{"user": "deploy-bot"}'
# Get, list and act by alias, as Opsgenie clients send them
curl "https://api.opsping.com/v2/alerts/db-1-disk-usage?identifierType=alias" \
-H "Authorization: GenieKey $OPSPING_API_KEY"
curl -G https://api.opsping.com/v2/alerts \
--data-urlencode "query=alias:db-1-disk-usage status:open" \
-H "Authorization: GenieKey $OPSPING_API_KEY"
curl -X POST "https://api.opsping.com/v2/alerts/db-1-disk-usage/close?identifierType=alias" \
-H "Authorization: GenieKey $OPSPING_API_KEY" \
-H "Content-Type: application/json" -d '{"user": "deploy-bot"}'
Heartbeats
curl -X POST https://api.opsping.com/v2/heartbeats/nightly-backup/ping \
-H "Authorization: GenieKey $OPSPING_API_KEY"
OpsPing answers with JSON, not Opsgenie's plain-text PONG, so update scripts that check the response body.
FAQ
Is OpsPing a drop-in replacement for the Opsgenie API?
For alert creation and the common alert actions, mostly yes: the same v2 paths, the Opsgenie JSON shape, alias deduplication, get-by-alias and the common actions by alias (identifierType=alias), alias: and status: tokens in list queries, Opsgenie-style assign, add-responder and tag-removal bodies, and the GenieKey header. It is not a full drop-in: unacknowledge, unassign, reopen, unsnooze, mark-read and delete still take the alert ID or tiny ID, there is no update-priority endpoint, request-status polling is not available, the rest of the Opsgenie query language is not implemented, responders must reference OpsPing objects, and configuration APIs use OpsPing-native schemas.
Which header do I use?
Send an OpsPing API key as Authorization: GenieKey <key>, Authorization: OpsPingKey <key>, or X-OpsPing-Key: <key>. All three take the new OpsPing key; your Opsgenie key will not work.
Will Alertmanager and Grafana auto-resolve alerts through the Opsgenie-compatible endpoint?
Yes. Both resolve Opsgenie alerts by calling close with identifierType=alias, which OpsPing supports, so keep send_resolved: true in Alertmanager and Auto close incidents (autoClose: true) in Grafana. OpsPing's native Alertmanager and Grafana webhook integrations remain an option if you want each alert mapped by its fingerprint.
What happens to responders in my existing payloads?
Responders must resolve to objects in your OpsPing tenant: users by ID or email username, and teams, schedules and escalations by ID or unambiguous name, subject to team access. Names must be unique within the caller’s accessible teams; missing or ambiguous names are rejected.
Is there an EU endpoint like api.eu.opsgenie.com?
Use https://api.opsping.com whether your Opsgenie account was in the US or EU region. Private single-tenant deployments use their own hostname.
OpsPing plans: Free (5 responders); Team $15/responder/month or $12 billed annually ($144/responder/year); Business $25/responder/month or $20 billed annually ($240/responder/year). Opsgenie switchers get 35% off the first annual term — 12 months, annual billing only — when redeemed by April 5, 2027 ($7.80/responder/month on Team, $13.00 on Business, first year).
Opsgenie® is a trademark of Atlassian. OpsPing is not affiliated with, sponsored, or endorsed by Atlassian.