AAP Prometheus Exporter

A standalone Prometheus exporter for Ansible Automation Platform (AAP) that exposes job metrics, platform resource counts, instance health, and license status by polling the AAP REST API (/api/controller/v2/).

Follows the Prometheus exporter best practices and the architectural pattern of windows_exporter: pluggable collectors, concurrent collection with per-collector timeouts, and CLI flag-driven cardinality controls.

---

Metrics

jobs collector

Metric	Type	Labels	Description
aap_jobs	Gauge	organization, job_template, status	All-time job count by template and status (successful, failed, canceled, error)
aap_jobs_running	Gauge	—	Currently running jobs
aap_jobs_pending	Gauge	—	Currently pending jobs
aap_job_duration_seconds	Histogram	organization, job_template	Job run durations within the lookback window

platform collector

Metric	Type	Labels	Description
aap_job_templates	Gauge	organization	Job template count per org
aap_workflow_job_templates	Gauge	organization	Workflow job template count per org
aap_projects	Gauge	organization	Project count per org
aap_inventories	Gauge	organization	Inventory count per org
aap_hosts	Gauge	organization	Host count per org
aap_schedules	Gauge	organization	Schedule count per org
aap_info	Gauge	version, license_type	Platform info (constant 1)
aap_license_expiry_seconds	Gauge	—	Seconds until license expiry
aap_license_instances_limit	Gauge	—	Maximum managed hosts allowed by license
aap_license_instances_free	Gauge	—	Free managed host slots remaining
aap_license_instances_current	Gauge	—	Currently managed hosts

instances collector

Metric	Type	Labels	Description
aap_instance_capacity	Gauge	hostname, node_type	Node capacity
aap_instance_consumed_capacity	Gauge	hostname, node_type	Node consumed capacity
aap_instance_cpu_cores	Gauge	hostname	CPU cores
aap_instance_memory_bytes	Gauge	hostname	Memory in bytes
aap_instance_info	Gauge	hostname, node_type, version, enabled	Node info (constant 1)

Orchestrator meta-metrics

Metric	Type	Labels	Description
aap_scrape_duration_seconds	Gauge	—	Total scrape duration
aap_collector_scrape_duration_seconds	Gauge	collector	Per-collector scrape duration
aap_collector_scrape_success	Gauge	collector	1 if collector succeeded, 0 otherwise
aap_collector_scrape_timeout	Gauge	collector	1 if collector hit the scrape timeout

---

Quick Start

./aap_exporter \
  --aap.url="https://your-aap-host/api/controller/v2/" \
  --aap.token="your-bearer-token"

Metrics available at http://localhost:9269/metrics.

Token from file

./aap_exporter \
  --aap.url="https://your-aap-host/api/controller/v2/" \
  --aap.token-file="/run/secrets/aap_token"

---

Configuration

Global flags

Flag	Env	Default	Description
--aap.url	AAP_URL	—	AAP API base URL (required, must end with /api/controller/v2/)
--aap.token	AAP_TOKEN	—	Bearer token
--aap.token-file	AAP_TOKEN_FILE	—	Path to token file
--aap.tls-skip-verify	—	false	Skip TLS certificate verification
--web.listen-address	—	:9269	Listen address
--web.telemetry-path	—	/metrics	Metrics path
--scrape.timeout	—	30s	Maximum scrape duration
--collectors.enabled	—	jobs,platform,instances	Comma-separated list of enabled collectors

jobs collector flags

Flag	Default	Description
--collector.jobs.include-templates	.+	Regexp of template names to include
--collector.jobs.exclude-templates	—	Regexp of template names to exclude
--collector.jobs.include-orgs	.+	Regexp of org names to include
--collector.jobs.exclude-orgs	—	Regexp of org names to exclude
--collector.jobs.concurrency	10	Parallel API calls for template fan-out
--collector.jobs.duration-lookback	24h	Lookback window for duration histogram

platform collector flags

Flag	Default	Description
--collector.platform.include-orgs	.+	Regexp of org names to include
--collector.platform.exclude-orgs	—	Regexp of org names to exclude

Cardinality example

Reduce jobs collector from ~700 API calls to ~24 by scoping to specific templates and orgs:

--collector.jobs.include-templates="^(deploy_prod|rollback|health_check)$" \
--collector.jobs.include-orgs="^(Ops|Platform)$"

---

Container

Docker — quick run

docker run --rm \
  -e AAP_URL="https://your-aap-host/api/controller/v2/" \
  -e AAP_TOKEN="your-bearer-token" \
  -p 9269:9269 \
  ghcr.io/dominik-esb/aap_exporter:latest

Metrics available at http://localhost:9269/metrics.

Docker — token from secret file

Mount a file and use --aap.token-file instead of the env var:

docker run --rm \
  -e AAP_URL="https://your-aap-host/api/controller/v2/" \
  -v /path/to/token.txt:/run/secrets/aap_token:ro \
  -p 9269:9269 \
  ghcr.io/dominik-esb/aap_exporter:latest \
  --aap.token-file=/run/secrets/aap_token

Docker Compose

Copy .env.example to .env, fill in your values, then:

docker compose up -d

docker-compose.yml reads AAP_URL and AAP_TOKEN from the environment (or .env). Uncomment the command lines to restrict which templates and orgs are scraped.

Build the image yourself

make docker-build              # docker build -t aap_exporter:latest .
make docker-build IMAGE=myrepo/aap_exporter TAG=1.0.0

Prometheus scrape config

scrape_configs:
  - job_name: aap
    static_configs:
      - targets: ["aap-exporter-host:9269"]
    scrape_interval: 60s
    scrape_timeout: 35s

Set scrape_timeout slightly above --scrape.timeout so Prometheus waits for the full scrape before declaring it stale.

---

Building from source

make build        # produces bin/aap_exporter
make test         # go test ./... -race -count=1
make lint         # go vet ./...

Requires Go 1.23+.

---

Project Layout

cmd/aap_exporter/        # Entry point
internal/
  client/                # AAP HTTP client (bearer auth, pagination)
  collector/             # Collector interface + orchestrator
  collector/jobs/        # jobs collector
  collector/platform/    # platform collector
  collector/instances/   # instances collector

---

License

MIT