Prometheus was created at SoundCloud in 2012 as a monitoring and alerting toolkit. It was inspired by Google’s Borgmon system, which was used internally at Google for monitoring large-scale distributed systems. The name “Prometheus” comes from Greek mythology, where Prometheus was a Titan who stole fire from the gods and gave it to humanity—symbolizing the project’s goal of bringing powerful monitoring capabilities to developers.
In 2012, SoundCloud engineers Matt T. Proud and Julius Volz began developing Prometheus as an open-source alternative to their existing monitoring solutions. The company was experiencing rapid growth and needed a monitoring system that could handle dynamic, service-oriented architectures.
Key motivations for creating Prometheus:
- Existing tools like Graphite were designed for static infrastructure
- Need for better support of multi-dimensional data (labels/tags)
- Desire for a pull-based model rather than push-based metrics
- Integration with service discovery for dynamic environments
The initial design drew heavily from Google’s Borgmon paper, adapting concepts like:
- Pull-based metric collection
- Multi-dimensional data model with labels
- PromQL (Prometheus Query Language) for data analysis
- Built-in alerting with Alertmanager
| Year |
Version |
Milestone |
| 2012 |
- |
Development begins at SoundCloud |
| 2013 |
0.x |
First internal releases at SoundCloud |
| 2014 |
0.x |
First public beta releases |
| 2015 |
- |
Prometheus joins CNCF as second hosted project |
| 2016 |
1.0 |
First stable release (July 2016) |
| 2016 |
1.1 |
Remote read/write API introduced |
| 2017 |
1.5 |
Improved service discovery |
| 2018 |
2.0 |
Major rewrite with new storage engine (November 2017) |
| 2018 |
2.1 |
Native histograms experimental |
| 2019 |
2.5 |
Thanos integration improvements |
| 2020 |
2.20 |
Enhanced remote write, federation improvements |
| 2021 |
2.30 |
Native histograms in beta |
| 2022 |
2.35 |
OTLP support, improved scrape protocols |
| 2023 |
2.45 |
Native histograms GA, improved scaling |
| 2023 |
3.0 |
Major architecture changes announced |
| 2024 |
3.0 |
Prometheus 3.0 officially released |
| 2025 |
3.5 |
LTS (Long Term Support) release |
| 2026 |
3.10.0 |
Current stable (February 2026) |
| 2026 |
3.5.x |
LTS branch |
The 1.x series established the core architecture:
- Time-series database with custom storage format
- Pull-based scraping over HTTP
- Multi-dimensional data model with metric names and labels
- PromQL for querying and aggregation
- Separate Alertmanager component
Version 2.0 was a complete rewrite with significant improvements:
-
New Storage Engine (TSDB)
- Built specifically for Prometheus workloads
- 40x more efficient than 1.x storage
- Native compression and compaction
- WAL (Write-Ahead Log) for crash recovery
-
Go-based UI
- Replaced Ruby/Sinatra UI with Go-based frontend
- Improved performance and easier deployment
- Modern expression browser
-
Service Discovery
- Native Kubernetes service discovery
- Consul, EC2, Azure, GCP integrations
- File-based service discovery
-
Remote Read/Write
- Long-term storage integration (Thanos, Cortex)
- Multi-cluster federation
- Cloud storage backends
Version 3.0 introduced groundbreaking changes:
-
OTLP Native Support
- OpenTelemetry Protocol integration
- Better interoperability with observability stack
- Reduced need for collectors
-
Enhanced Scalability
- Improved sharding capabilities
- Better resource utilization
- Native support for high-cardinality metrics
-
Native Histograms GA
- Exponential bucketing
- Reduced storage requirements
- Better percentile calculations
-
PromQL Enhancements
- New functions for advanced analysis
- Improved subquery support
- Better performance optimization
Prometheus was the second project to be hosted by the Cloud Native Computing Foundation (CNCF), after Kubernetes. This was a pivotal moment:
- March 2015: Initial CNCF application
- May 2015: Accepted as incubating project
- August 2016: Graduated to CNCF graduated project
Being part of CNCF provided:
- Legal framework for contributions
- Vendor-neutral governance
- Marketing and community support
- Integration with Kubernetes ecosystem
- Access to CNCF events and promotion
- 2016: CNCF Incubating → Graduated
- One of the fastest projects to graduate
- Demonstrated strong adoption and community
The Prometheus ecosystem grew rapidly with community-contributed exporters:
- Official Exporters: node_exporter, blackbox_exporter, pushgateway
- Database Exporters: mysqld_exporter, postgres_exporter, redis_exporter
- Application Exporters: jmx_exporter, php-fpm_exporter
- Cloud Exporters: cloudwatch_exporter, stackdriver_exporter
- Hardware Exporters: ipmi_exporter, smartctl_exporter
Major platforms integrated Prometheus support:
- Kubernetes: Native metrics endpoint
- Docker: Stats API with Prometheus format
- AWS: Managed Service for Prometheus
- Google Cloud: Managed Prometheus
- Azure: Azure Monitor for Prometheus
- Red Hat: OpenShift Metrics with Prometheus
The ecosystem spawned several major projects:
- Thanos (2018): Long-term storage and global querying
- Cortex (2017): Multi-tenant, horizontally scalable Prometheus
- VictoriaMetrics (2019): High-performance alternative storage
- Mimir (2022): Grafana Labs’ distributed Prometheus
- Prometheus Operator (2017): Kubernetes-native deployment
In 2020, Grafana Labs hired several core Prometheus maintainers:
- Björn Rabenstein (Senior Software Engineer)
- Julius Volz (Co-founder, Prometheus creator)
- Other key contributors joined over time
This was not an acquisition of the project (which remains CNCF-owned) but strategic hiring to ensure continued development.
Major companies investing in Prometheus development:
- Grafana Labs: Primary sponsor, employs core maintainers
- Red Hat: OpenShift integration
- Google: Borgmon heritage, GKE integration
- AWS: Managed Prometheus service
- DigitalOcean: Managed Prometheus
- Bloomberg: Large-scale deployments
- GitHub Stars: 55,000+
- Contributors: 800+
- Downloads: Millions monthly
- CNCF Status: Graduated Project
- Adoption: Industry standard for cloud-native monitoring
- Regular monthly releases
- Active security patching
- Growing exporter ecosystem
- Strong enterprise adoption
- Enhanced OTLP Support: Deeper OpenTelemetry integration
- Improved Scalability: Better handling of massive deployments
- AI/ML Integration: Anomaly detection, forecasting
- Edge Computing: Lightweight deployments for IoT/edge
- Security Enhancements: mTLS, improved authentication
Any questions?
Feel free to contact us. Find all contact information on our contact page.