Send Your First Telemetry

InfraSage accepts telemetry in three formats. This guide shows you how to send each type and verify the data was received.

Format 1: plain JSON (recommended for getting started)

The simplest way to send a single metric:

curl -X POST $INFRASAGE_URL/api/v1/telemetry \
  -H "Content-Type: application/json" \
  -d '{
    "service_id": "payment-api",
    "metric_name": "request_latency_ms",
    "value": 142.5,
    "timestamp": "2026-04-10T12:00:00Z",
    "attributes": {
      "region": "us-east-1",
      "env": "production",
      "version": "v2.3.1"
    }
  }'

Batch ingestion (up to 10,000 records per request)

curl -X POST $INFRASAGE_URL/api/v1/telemetry/batch \
  -H "Content-Type: application/json" \
  -d '[
    {
      "service_id": "payment-api",
      "metric_name": "request_latency_ms",
      "value": 142.5,
      "timestamp": "2026-04-10T12:00:00Z"
    },
    {
      "service_id": "payment-api",
      "metric_name": "error_rate",
      "value": 0.02,
      "timestamp": "2026-04-10T12:00:00Z"
    },
    {
      "service_id": "user-service",
      "metric_name": "cpu_usage_percent",
      "value": 72.1,
      "timestamp": "2026-04-10T12:00:00Z"
    }
  ]'

Format 2: OpenTelemetry (OTLP)

Configure your OTEL collector to export to InfraSage:

# otel-collector-config.yaml
exporters:
  otlphttp:
    endpoint: $INFRASAGE_URL
    headers:
      Content-Type: application/x-protobuf

service:
  pipelines:
    metrics:
      exporters: [otlphttp]
    traces:
      exporters: [otlphttp]
    logs:
      exporters: [otlphttp]

Or send raw OTLP JSON:

curl -X POST $INFRASAGE_URL/v1/metrics \
  -H "Content-Type: application/json" \
  -d '{
    "resourceMetrics": [{
      "resource": {
        "attributes": [{"key": "service.name", "value": {"stringValue": "checkout-service"}}]
      },
      "scopeMetrics": [{
        "metrics": [{
          "name": "http.request.duration",
          "gauge": {
            "dataPoints": [{
              "asDouble": 0.125,
              "timeUnixNano": "1712750400000000000"
            }]
          }
        }]
      }]
    }]
  }'

Format 3: Prometheus remote write

Add InfraSage as a remote-write target in your prometheus.yml:

remote_write:
  - url: $INFRASAGE_URL/api/v1/prometheus/write
    queue_config:
      max_samples_per_send: 10000
      batch_send_deadline: 5s

Telemetry types

InfraSage supports six telemetry types. Specify the type in the type field (defaults to metric):

Type	hasValue	hasBody	hasTrace	Use Case
metric	Yes	No	No	CPU, memory, latency, throughput
log	No	Yes	No	Application logs
trace	Yes	No	Yes	Distributed tracing, spans
event	No	Yes	No	Kubernetes events, deploys
profile	Yes	Yes	No	CPU/memory profiling
slo	Yes	No	No	SLI/SLO compliance tracking

Sending a log entry

curl -X POST $INFRASAGE_URL/api/v1/telemetry \
  -H "Content-Type: application/json" \
  -d '{
    "service_id": "auth-service",
    "type": "log",
    "body": "User login failed: invalid credentials",
    "timestamp": "2026-04-10T12:00:00Z",
    "attributes": {
      "level": "warn",
      "user_id": "u-12345"
    }
  }'

Sending a trace span

curl -X POST $INFRASAGE_URL/api/v1/telemetry \
  -H "Content-Type: application/json" \
  -d '{
    "service_id": "checkout-service",
    "type": "trace",
    "metric_name": "checkout.duration",
    "value": 0.342,
    "trace_id": "abc123def456",
    "timestamp": "2026-04-10T12:00:00Z"
  }'

Sending an SLO reading

curl -X POST $INFRASAGE_URL/api/v1/telemetry \
  -H "Content-Type: application/json" \
  -d '{
    "service_id": "api-gateway",
    "type": "slo",
    "metric_name": "availability",
    "value": 99.97,
    "timestamp": "2026-04-10T12:00:00Z"
  }'

Verify the data arrived

Check the ingestion counter

curl -s '$PROMETHEUS_URL/api/v1/query?query=infrasage_ingestion_events_total' | \
  jq '.data.result[].value[1]'

See your services

Open Services in the console: each service you sent data for appears within a minute or two. Or ask the API:

curl $INFRASAGE_URL/api/v1/dashboards/services \
  -H "Authorization: Bearer $YOUR_JWT"

List supported telemetry types via API

curl $INFRASAGE_URL/api/v1/telemetry-types

Validation rules

InfraSage validates every record before queueing. Records that fail validation go to the dead-letter queue (DLQ) rather than being silently dropped.

Rule	Details
service_id required	Must be non-empty
timestamp age	Must be within ±24 hours of current time
value range	Must be a finite float64 (no NaN, no Inf)
metric_name length	Maximum 256 characters
type enum	Must be one of: metric, log, trace, event, profile, slo

Failed records are stored in the DLQ with their original payload for manual inspection and replay:

curl $INFRASAGE_URL/api/v1/debug/dlq-stats