Services API

A service is one service.name that sends server or consumer spans. The Services API returns the request rate, error rate, latency, and health of each service, its operations and dependencies, and the map of calls between services and databases. See Services and releases for how Squasher finds services and sets their health.

Operations

Operation	Method and path	Purpose
`services.list`	`GET /v1/projects/{project_id}/services/catalog`	List services with metrics, health, and the last release.
`services.get`	`GET /v1/projects/{project_id}/services/catalog/{service_name}`	Read one service: summary, series, operations, dependencies, and releases.
`services.map`	`GET /v1/projects/{project_id}/services/map`	Read the service map: service and database nodes, and the calls between.

All three operations need the errors:read scope. URL-encode service_name.

Query parameters

Every operation takes a window and an environment:

Parameter	Values	Default
`from`	ISO 8601 timestamp	24 hours before `to`
`to`	ISO 8601 timestamp	Now
`environment`	A `deployment.environment.name`, for example `production`. Empty (`environment=`): telemetry without an environment	All environments

The window is at most 7 days. services.list also takes these parameters:

Parameter	Values	Default
`health`	`healthy`, `degraded`, `failing`, `no_data`	All services
`q`	Text that the service name contains, at most 200 characters	All services
`sort`	`count`, `error_rate`, `p95`, `last_seen`	`count` (most requests first)
`limit`	1 to 100	100

List services

curl --get \
  --url "https://api.squasher.ai/v1/projects/$SQUASHER_PROJECT_ID/services/catalog" \
  --header "x-squasher-key: $SQUASHER_API_KEY" \
  --data-urlencode "environment=production" \
  --data-urlencode "sort=error_rate"

{
  "data": [
    {
      "service_name": "checkout-api",
      "status": "degraded",
      "request_count": 48210,
      "error_count": 731,
      "error_rate": 1.52,
      "p50_ms": 41,
      "p95_ms": 380,
      "p99_ms": 1210,
      "throughput_per_min": 33.5,
      "last_seen": "2026-10-11T09:29:41.000Z",
      "last_release": {
        "id": "rel_WyJwcm9kdWN0aW9uIiwiY2hlY2tvdXQtYXBpIiwiMi40LjEiXQ",
        "service_name": "checkout-api",
        "environment": "production",
        "version": "2.4.1",
        "confidence": "confirmed",
        "started_at": "2026-10-10T16:02:11.000Z",
        "ended_at": null,
        "is_current": true,
        "first_seen": "2026-10-10T16:02:40.000Z",
        "last_seen": "2026-10-11T09:29:41.000Z",
        "commit_sha": "9f2c1e7",
        "deployment_count": 1,
        "previous_release_id": "rel_WyJwcm9kdWN0aW9uIiwiY2hlY2tvdXQtYXBpIiwiMi40LjAiXQ"
      }
    }
  ],
  "status_counts": { "healthy": 6, "degraded": 1, "failing": 0, "no_data": 1 },
  "total_services": 8,
  "environment": "production",
  "environments": ["production", "staging"],
  "from": "2026-10-10T09:30:00.000Z",
  "to": "2026-10-11T09:30:00.000Z",
  "truncated": false
}

The JSON is shortened to one service.

Field	Description
`data`	One row for each service: metrics, `status` (health), `last_seen`, and `last_release` (`null` without a release).
`counts_approximate`	`true` when the project has more services than one read groups (about 20,000). `status_counts` and `total_services` then count only part of them.
`status_counts`	Services for each health, over every service in the window, before the `health` filter and `limit`.
`total_services`	Services that match `environment`, `q`, and `health`, before `limit`.
`environments`	Environments that sent telemetry in the window. Use one as `environment`.
`truncated`	`true` when more services match than the response holds.

Rates are percents from 0 to 100. Latencies are in milliseconds. throughput_per_min is requests per minute over the window.

Read one service

curl --get \
  --url "https://api.squasher.ai/v1/projects/$SQUASHER_PROJECT_ID/services/catalog/checkout-api" \
  --header "x-squasher-key: $SQUASHER_API_KEY" \
  --data-urlencode "environment=production"

{
  "service_name": "checkout-api",
  "status": "degraded",
  "environment": "production",
  "environments": ["production", "staging"],
  "from": "2026-10-10T09:30:00.000Z",
  "to": "2026-10-11T09:30:00.000Z",
  "granularity": "hour",
  "summary": {
    "request_count": 48210,
    "error_count": 731,
    "error_rate": 1.52,
    "p50_ms": 41,
    "p95_ms": 380,
    "p99_ms": 1210,
    "throughput_per_min": 33.5,
    "last_seen": "2026-10-11T09:29:41.000Z"
  },
  "series": [
    {
      "time": "2026-10-10T10:00:00.000Z",
      "request_count": 2011,
      "error_count": 22,
      "error_rate": 1.09,
      "p50_ms": 40,
      "p95_ms": 352,
      "p99_ms": 1104
    }
  ],
  "operations": [
    {
      "operation_name": "POST /checkout",
      "request_count": 12904,
      "error_count": 402,
      "error_rate": 3.12,
      "p50_ms": 95,
      "p95_ms": 610,
      "p99_ms": 1820,
      "last_seen": "2026-10-11T09:29:41.000Z"
    }
  ],
  "dependencies": {
    "upstream": [
      { "name": "web", "request_count": 40112, "error_rate": 1.4, "avg_latency_ms": 88 }
    ],
    "downstream": [
      { "name": "payments", "request_count": 12790, "error_rate": 0.6, "avg_latency_ms": 140 }
    ],
    "databases": [
      { "name": "postgresql", "request_count": 96044, "error_rate": 0, "avg_latency_ms": 4.2 }
    ]
  },
  "releases": []
}

upstream lists the services that call this service. downstream lists the services it calls. databases lists the database systems it calls, by db.system.name.
releases lists the releases of the service that served in the window, newest first. Each item has the same fields as last_release above.
series has one point for each granularity bucket.

A service that sent no spans in the window answers with status: "no_data" and empty sections, not an error.

Read the service map

curl --get \
  --url "https://api.squasher.ai/v1/projects/$SQUASHER_PROJECT_ID/services/map" \
  --header "x-squasher-key: $SQUASHER_API_KEY" \
  --data-urlencode "environment=production"

{
  "nodes": [
    {
      "id": "service:checkout-api",
      "kind": "service",
      "name": "checkout-api",
      "status": "degraded",
      "request_count": 48210,
      "error_rate": 1.52,
      "avg_latency_ms": 96
    },
    {
      "id": "database:postgresql",
      "kind": "database",
      "name": "postgresql",
      "status": "healthy",
      "request_count": 96044,
      "error_rate": 0,
      "avg_latency_ms": 4.2
    }
  ],
  "edges": [
    {
      "source": "service:checkout-api",
      "target": "database:postgresql",
      "kind": "database",
      "request_count": 96044,
      "error_rate": 0,
      "avg_latency_ms": 4.2
    }
  ],
  "environment": "production",
  "environments": ["production", "staging"],
  "from": "2026-10-10T09:30:00.000Z",
  "to": "2026-10-11T09:30:00.000Z"
}

A node id is service:<service.name> or database:<db.system.name>. An edge kind is call (service to service) or database (service to database). source and target are node IDs.

Errors

Status	When
`400`	A parameter is not valid, for example an unknown `health` or a `limit` above 100.
`403`	The key does not have `errors:read` on the project.
`422`	`from` is after `to`, or the window is longer than 7 days.

TypeScript client, CLI, and MCP

const services = await client.services.list({ environment: "production", health: "failing" });
const checkout = await client.services.get("checkout-api", { environment: "production" });
const map = await client.services.map({ environment: "production" });

squasher services list --project <project_id> --environment production --sort error_rate
squasher services get --project <project_id> checkout-api
squasher services map --project <project_id> --environment production

Add --json to print the full response. In MCP Code Mode, call search("services") first, then execute with services.list, services.get, or services.map.

Start here

Resources

Operations

Query parameters

List services

Read one service

Read the service map

Errors

TypeScript client, CLI, and MCP

​Operations

​Query parameters

​List services

​Read one service

​Read the service map

​Errors

​TypeScript client, CLI, and MCP

​Related guides

Operations

Query parameters

List services

Read one service

Read the service map

Errors

TypeScript client, CLI, and MCP

Related guides