Apica Forge
Time Series Database
Where Your Metrics Become the Foundation for Agentic Intelligence
Store Billions of Metrics. Query Instantly. Never Compromise.
Flow governs the data. Lake stores the telemetry. Forge powers the metrics.
While Lake stores all observability data at object storage economics, Forge is purpose-built for one thing legacy platforms can’t handle: Extreme-scale, high-cardinality metrics at machine speed. A distributed time-series database engineered for resiliency, simplistic operation, and embedded analytics, Forge stores billions of unique metric streams, token costs, agent latencies, SLI/SLO signals, multi-agent call rates, without performance degradation or cost explosion.
Because Forge is distributed and continuously replicating, your metrics remain available and queryable even during node failures or maintenance windows. No gaps in agent telemetry. No alerting blind spots. No capacity planning. Just the metrics backbone your agentic infrastructure depends on.
The Problem
Is Limited Visibility, Data Volume, and Query Performance Hindering Your IT Monitoring?
Traditional time series databases weren’t built for modern cloud-native cardinality demands
Cardinality Pricing Traps
Competing platforms tie cardinality limits to pricing tiers, forcing impossible trade-offs between monitoring granularity and budget.
Query Performance Degradation
Traditional TSDBs slow to a crawl under high-cardinality workloads — returning results in minutes for queries that should take milliseconds.
Exponential Cost Scaling
Per-metric and per-custom-metric pricing models penalize cloud-native architectures, with costs growing 250% year-over-year as infrastructure scales.
Premature Aggregation
Teams forced to aggregate data at collection time destroy granularity and create blind spots that prevent accurate root cause analysis.
Tag Anxiety
Engineers self-censor meaningful metric labels to stay within cardinality budgets, reducing the observability value of every data point collected.
Vendor Lock-In
Proprietary query languages, storage formats, and integration requirements make switching painful and create long-term dependency risks.
Capabilities
What Forge Does
A production-validated TSDB engineered for unlimited cardinality, millisecond query performance, and cloud-native scale.
High-Cardinality Metrics at Petabyte Scale
Store and query billions of unique metric streams, including the AI-specific signals that break general-purpose TSDBs: Per-token costs, per-agent latency histograms, tool invocation rates, and multi-agent call chain metrics, without performance penalties or cost explosions as cardinality grows.
Continuous Operations Under Failure
Multi-node replication distributes metric data across the cluster so Forge keeps ingesting and serving queries even when individual nodes go offline. When nodes recover, they automatically catch up from peer journals, no operator intervention, no data loss.
Embedded Analytics and Computation
Turn raw agent telemetry into actionable intelligence directly in the database layer: Establish reliable baselines, detect anomalies, calculate cost attribution, and derive SLI/SLO thresholds before data ever reaches a visualization or alerting platform.
Open Ecosystem, Zero Lock-In
Native support for Graphite, OpenTSDB, and Prometheus ingestion formats, with a purpose-built Grafana data source integration. Forge works alongside your existing metrics tooling and scales as your agentic footprint grows, no proprietary formats, no vendor dependency.
Purpose-Built for Modern Infrastructure Scale
Designed for installations managing millions of time series with high availability requirements, Forge handles the data surges, outages, and routine maintenance that defeat traditional TSDBs, making it the right foundation for organizations running AI workloads, Kubernetes environments, and distributed microservices at enterprise scale.
Benefits
Why Forge
From eliminating cardinality traps to millisecond query performance — built for modern cloud-native monitoring.
Unlimited Cardinality Without Cost Explosion
No per-metric pricing penalties. Store Hundreds of millions of unique metric streams with consistent performance and predictable costs.
Millisecond Query Performance at Scale
Forge maintains millisecond-level query responses under high-cardinality workloads that bring traditional TSDBs to their knees.
Enhanced Monitoring Visibility
Reliable, scalable, and fast — improving visibility across IT operations without the blind spots caused by premature aggregation.
Seamless Data Integration
Ingests various data sources and integrates with existing monitoring systems — no proprietary formats that create lock-in.
High Availability by Design
Data replication and clustering keep telemetry data available even during outages — no single points of failure.
Tag Freely
Eliminate tag anxiety — instrument with as many labels as your observability needs demand without budget or performance consequences.
Cloud-Native Architecture
Built for the scale and dynamism of cloud-native infrastructure — not retrofitted from an on-premises era design.
Open Integrations
Works with Graphite, OpenTelemetry, Prometheus, and other open standards — protecting your monitoring investment long-term.
Proven Business Outcomes
Production-Validated at Enterprise Scale
Apica Forge is not a new product. It has been battle-tested in production environments handling some of the most demanding cardinality workloads in cloud-native infrastructure
Case Study
Enterprise Cloud-Native Platform
Large-scale Kubernetes deployment generating tens of millions of unique metric streams. Existing platform imposing cardinality limits and forcing tag dimensionality compromises that degraded incident response.
Results:
- 10–15× increase in metric cardinality without performance impact
- Eliminated all tag dimensionality compromises — every label retained
- 60–70% reduction in metrics infrastructure costs
- Engineers no longer self-censoring labels to stay within cardinality budgets
Case Study
Multi-Cloud Enterprise (AWS/Azure/GCP)
Hybrid multi-cloud infrastructure with Istio/Envoy service mesh generating extreme tag cardinality. Prometheus federation approach couldn’t scale past 30-day retention across 200+ instances.
Results:
- Centralized metrics from 200+ Prometheus instances across 3 cloud providers
- Extended data retention from 30 days to 2 years without storage cost explosion
- Enabled cross-cloud capacity planning and year-over-year trend analysis
- Sub-100ms query latency maintained across entire multi-cloud footprint
FAQ
Frequently Asked Questions
How does Apica Forge handle high-cardinality workloads differently than traditional TSDBs?
Traditional TSDBs were designed for relatively low cardinality and bolt on high-cardinality support as an afterthought — imposing artificial limits, degrading query performance, or charging exponential overages. Apica Forge’s tag-first indexing engine and distributed architecture were designed from the ground up for Hundreds of millions of unique metric streams. This means you get consistent sub-100ms query latency and no cardinality caps, even as your Kubernetes pods, microservices, and multi-cloud dimensions multiply.
What does "histogram-native storage" mean in practice?
Instead of storing every individual measurement (raw samples), Apica Forge natively stores data in histogram buckets. This enables accurate P50, P95, P99, and P99.9 percentile calculations without requiring you to retain millions of raw data points. The result is 10–100× storage reduction while maintaining the statistical accuracy your SREs need for latency analysis, SLO calculations, and anomaly detection.
Can Apica Forge replace our existing Prometheus setup?
Yes. Apica Forge can serve as a long-term storage backend for Prometheus, replacing the 30-day retention limit that most teams hit as their infrastructure scales. One multi-cloud enterprise used Apica Forge to centralize 200+ Prometheus instances across AWS, Azure, and GCP — extending retention from 30 days to 2 years while maintaining sub-100ms query latency across their entire footprint. Your existing Prometheus configuration, exporters, and PromQL tooling continue to work.
Is Apica Forge available on-premises, in the cloud, or both?
Apica Forge supports flexible deployment models including on-premises, cloud, and multi-datacenter hybrid configurations. The distributed architecture allows you to deploy across availability zones and cloud providers simultaneously. Fast deployment means installation takes minutes with a single command, and scaling is a single operation — no complex cluster management required.
What does Apica Forge's pricing model look like?
Apica Forge uses consumption-aligned, capacity-based pricing agreements — not per-metric or per-custom-metric pricing. This means your costs don’t grow exponentially as your infrastructure scales. The capacity-based model provides cost predictability while supporting substantial infrastructure growth within contracted parameters. Contact your Apica representative for a customized quote based on your specific requirements.
Does Apica Forge integrate with Grafana and our existing dashboards?
Yes. Apica Forge ships with a native Grafana data source plugin that allows you to expand your metrics capacity without changing your visualization stack. It also provides full Graphite-web compatibility (0.10 and later), making it a drop-in solution for organizations already running Graphite. Your existing dashboards, alerts, and queries continue to function — Apica Forge simply removes the cardinality and cost constraints.
Integrations
Works With Your Existing Stack
Forge integrates with Graphite, OpenTelemetry, Prometheus, and the full range of metrics collection tools.
Graphite
OpenTelemetry
Prometheus
Telegraf
Grafana
StatsD
collectd
Kubernetes
Don’t see your metrics tool? Forge integrates with any metrics collection tool that supports open standards — no proprietary formats required.