dt-obs-ext-monitors
>-
Works with
Claude CodeCursorCodex CLIGitHub CopilotGemini CLI
---
name: dt-obs-ext-monitors
description: >-
license: Apache-2.0
---
# External Monitor Ingestion
Send 3rd-party test and monitor results to Dynatrace Grail using the events ingest API.
This is the canonical replacement for the deprecated `POST /api/v1/synthetic/ext/tests` endpoint.
## Overview
Events posted to `/platform/ingest/custom/events/{endpoint}` land in Grail and are processed
by OpenPipeline, which:
- Extracts metrics (`external.test.availability`, `external.test.duration`) for alerting and SLOs
- Registers each unique `test.id` as an `EXT_TEST` Smartscape node — enabling Davis Problems to
attach to a named entity ("External test X went down") rather than floating without topology context
- Adds `result.status.category` to step events (`SUCCESS` / `SKIPPED` / `FAIL`) for dashboard filtering
Two event types form a test result:
| Type | Purpose |
|------|---------|
| `external_test_run` | Overall pass/fail result for one test execution |
| `external_test_step` | One step within that run (optional; enables step-level metrics) |
## Authentication
Two token types are accepted:
| Token type | Scope |
|------------|-------|
| Classic Api-Token | `openpipeline.events.custom` |
| Platform Token / OAuth | `openpipeline:events.custom:ingest` |
Common wrong guess that does NOT work: `events.ingest`.
## Quick Start
> **Prerequisite:** The ingest endpoint must be created in OpenPipeline before sending events.
> The `external.tests` endpoint is provisioned automatically by the
> default Dynatrace 3rd-party monitors Monaco bundle.
> See `references/event-ingestion.md` for setup details and custom endpoint creation.
Send one minimal test result (replace `external.tests` with your configured endpoint name):
```text
curl -X POST "https://{env-id}.live.dynatrace.com/platform/ingest/custom/events/external.tests" \
-H "Authorization: Api-Token {token}" \
-H "Content-Type: application/json" \
-d '[{
"event.kind": "EXTERNAL_TEST_EVENT",
"event.type": "external_test_run",
"test.id": "my-api-health-check",
"test.run.id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
"test.name": "My API Health Check",
"test.type": "api",
"test.run.status": "passed",
"test.run.availability": 1,
"test.run.duration_ms": 245,
"test.run.location": "us-east-1",
"dt.security_context": "team-checkout",
"timestamp": "2024-01-15T10:30:00Z"
}]'
```
**Expected response:** HTTP 200 (empty body).
> Full event schema, step events, extended examples with error fields and CI metadata,
> Java DTO shapes, and DQL queries: `references/event-ingestion.md`
## Verify in Grail
After sending, confirm the event appears:
```dql
fetch events, from:now()-1h
| filter event.type == "external_test_run"
| fields timestamp, test.id, test.name, test.run.status, test.run.duration_ms, test.run.location
| sort timestamp desc
| limit 20
```
## Key Constraints
- Body **must** be a JSON array (`[{...}]`), not an object wrapper (`{"events":[...]}`).
- A single request can mix run events and step events in the same array.
- `timestamp` must be ISO 8601 UTC (e.g. `"2024-01-15T10:30:00Z"`).
- `test.run.availability` must be integer `1` or `0`, not a string.
- All step events in a run must share the same `test.run.id` and `test.id` as the parent run event.
- Step events must include `test.step.id` (unique within the run) in addition to `test.run.id`.
- `event.kind: "EXTERNAL_TEST_EVENT"` must be present — used by Grail for event classification.
Routing to the correct pipeline is driven by `event.type`, not `event.kind`.
- Do **not** send `dt.smartscape.ext_test` or `result.status.category` — these are written by
the pipeline after ingestion and will be overwritten if included.
- `dt.security_context` controls data access policies; it’s recommended to set it on every event to enable
per-team access control and cost attribution in multi-team tenants.
## Multi-Location Tests
Send the same `test.id` from multiple locations — each with a different `test.run.location`
value — to build a multi-location test. OpenPipeline creates one `EXT_TEST` node per `test.id`
and one availability metric timeseries per `(test.id, location)` pair.
This enables two alerting tiers out of the box:
| Alert type | Fires when |
|------------|-----------|
| Local outage | A single location's availability drops (per-location timeseries) |
| Global outage | Average across all locations drops below threshold (e.g. majority failing) |
Threshold maths for a 3-location test: 1 location failing → avg 0.67 (no alert at 0.5 threshold);
2 failing → avg 0.33 (fires); all 3 failing → avg ≈ 0 (fires immediately).
Keep `test.id` stable across all locations — changing it creates a new Smartscape node and
breaks metric history.
## Related Skills
- **dt-dql-essentials** — DQL syntax for querying ingested events and building analysis queries
- **dt-obs-frontends** — Link test runs to frontend entities via `dt.smartscape.frontend` to
draw EXT_TEST → FRONTEND dependency edges in SmartscapeMore General & Other skills
find-skills
vercel-labs/skills
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.
1.5M
grill-me
mattpocock/skills
A relentless interview to sharpen a plan or design.
972.7k
grill-with-docs
mattpocock/skills
A relentless interview to sharpen a plan or design, which also creates docs (ADR's and glossary) as we go.
828.8k

