Welcome to PowerPlans Grid

PowerPlans Grid is a managed metrics platform for infrastructure teams that need consistent, long-tail visibility across thousands of hosts, hundreds of services and dozens of regions — without running Prometheus and its trailing stack of exporters, sidecars and storage layers.

Quickstart

Install the agent, send your first metric and query it from the dashboard in under five minutes.

Architecture

How agents, the ingest fleet, the time-series store and the query gateway fit together.

API Reference

REST + gRPC reference for the /v1 API — endpoints, auth, error codes.

Install agents

Packaged for Linux (deb/rpm), Kubernetes (Helm), FreeBSD, Windows Server and OpenBSD.

Quickstart

If you have a workspace already, this takes about four minutes. If you don't, create one first (14-day trial, no card).

1. Install the agent

# Debian / Ubuntu
curl -fsSL https://pkg.powerplans.org/install.sh | sudo bash -s -- \
  --workspace="$POWERPLANS_WORKSPACE" \
  --api-key="$POWERPLANS_API_KEY"

2. Confirm the handshake

Within 30 seconds the node appears under Grid → Nodes. Heartbeats arrive every 15 s by default.

3. Send a custom metric

curl -X POST https://api.powerplans.org/v1/metrics/ingest \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "metric":"orders.completed",
    "tags": {"region":"eu-west","service":"checkout"},
    "points":[{ "t":"2026-11-01T12:00:00Z", "v":1 }]
  }'

Architecture

Grid separates ingest, storage and query into three independently-scaled fleets. Agents talk gRPC to the ingest fleet; the dashboard and SDK clients talk REST to the query gateway.

┌──────────────┐ gRPC ┌──────────────┐ write ┌──────────────┐ │ Agents │ ───────────▶ │ Ingest │ ─────────────▶ │ Hot Store │ │ (Linux, K8s, │ │ Fleet │ │ (last 14d, │ │ FreeBSD, …) │ ◀────── ack │ (5× regions) │ │ NVMe, 1s) │ └──────────────┘ └──────────────┘ └──────┬───────┘ │ roll-up ▼ ┌──────────────┐ HTTPS ┌──────────────┐ read ┌──────────────┐ │ Dashboard │ ◀─────────── │ Query │ ◀──────────── │ Warm Store │ │ / SDK │ │ Gateway │ │ (365d, S3) │ └──────────────┘ └──────────────┘ └──────────────┘

Ingest fleet accepts bidi-streamed metric points over gRPC. Each ingest pod handles up to 120K points/s.
Hot store holds 14 days at full resolution on NVMe.
Warm store holds up to 365 days at 60-second resolution in object storage (per-region).
Query gateway federates queries across hot + warm with predicate pushdown and cost estimation.

Grid & Nodes

A node is any process running the PowerPlans agent, identified by its node_id (stable across restarts). Nodes belong to one workspace and can be labelled with free-form tags (region, env, service, …) used for query filtering.

A grid is a logical group of nodes; grids let you apply shared alerting, retention and RBAC policies.

Metrics model

PowerPlans uses a multi-dimensional metrics model close to Prometheus's:

Counter — monotonically increasing. Reset detection is automatic.
Gauge — point-in-time value, up or down.
Histogram — pre-aggregated bucketed distribution.
Summary — client-side quantile estimates.

Series cardinality limit per workspace: 2M active series on Free, 25M on Business, unlimited on Enterprise (with soft advisory at 250M).

Retention & roll-ups

Tier	Resolution	Retention	Backed by
Hot	1 second	14 days	NVMe (3× replicated)
Warm	60 seconds	365 days	Object storage (3× regions)
Cold (archive)	5 minutes	7 years	Glacier-class storage

Storage tiers

You never move data between tiers manually — the ingest pipeline promotes points through the tiers on a configurable schedule. For compliance (e.g. EU-only) you can pin a workspace to a specific region; cross-region replication can be disabled in Enterprise plans.

Install the agent

Packages are signed with the pkg.powerplans.org GPG key (fingerprint 5A17 9D47 … 8E22). Verify before installing.

Debian / Ubuntu

curl -fsSL https://pkg.powerplans.org/gpg | sudo gpg --dearmor -o /usr/share/keyrings/powerplans.gpg
echo "deb [signed-by=/usr/share/keyrings/powerplans.gpg] https://pkg.powerplans.org/deb stable main" | \
  sudo tee /etc/apt/sources.list.d/powerplans.list
sudo apt update
sudo apt install powerplans-agent

Kubernetes (Helm)

helm repo add powerplans https://charts.powerplans.org
helm install pp-agent powerplans/agent \
  --set workspace=$POWERPLANS_WORKSPACE \
  --set apiKey=$POWERPLANS_API_KEY

Monitoring Kubernetes

The Helm chart deploys the agent as a DaemonSet, scrapes the kubelet, api-server and node-exporter, and tags all metrics with cluster, namespace, pod and container. A default dashboard is auto-provisioned under Dashboards → Kubernetes.

Custom metrics

Three ways to emit custom metrics, in order of preference:

Prometheus endpoint on the app — the agent scrapes it on the loopback.
StatsD / DogStatsD to 127.0.0.1:8125 — the agent aggregates and forwards.
REST ingest — POST /v1/metrics/ingest, up to 5K points per request.

Alerting

Alerts are declared in YAML and evaluated server-side every 30 s. Destinations: email, Slack, Microsoft Teams, PagerDuty, Opsgenie, generic webhook.

- name: high-ingest-latency
  expr: rate(ingest_latency_p95[5m]) > 200ms
  for: 10m
  severity: page
  destinations: [pagerduty-primary]

Security

All traffic (agent → ingest, SDK → gateway, dashboard → gateway) is TLS 1.2+ with ECDSA certificates rotated every 60 days. Agent tokens are per-node, scoped to ingest-only, and rotated on heartbeat if the server requests rotation.

Data protection

EU data residency option (EU-only storage and processing)
Encryption in transit (TLS 1.2+) and at rest (AES-256)
Per-workspace audit log with 12-month retention
Data export and deletion on request

Regions

eu-west-1 — Warsaw, Poland (primary)
eu-central-1 — Frankfurt, Germany
eu-north-1 — Stockholm, Sweden
us-east-1 — Ashburn, Virginia
us-west-2 — Portland, Oregon
ap-southeast-1 — Singapore

Workspaces are pinned to one primary region. Cross-region failover is included on Business and Enterprise plans.