Backstory:
We spoke to dozens of observability practitioners in large enterprises. Across the board, we heard two things:
(1)The RoI on observability tools is hard to justify. (2)OSS leads to tool sprawl, metrics in Grafana, logs in ELK, traces… somewhere.
We wanted a simple, unified, and fast observability platform that you can deploy on your own infra, whether it’s bare metal, VMs, or Kubernetes. Prism is the UI layer of that platform.
What it does: (1)Fast Log Explorer with real-time ingestion graphs, one-click filters, saved views, and auto-detected formats. (2)Sub-second Text Search using structured, on-demand indexing over logs stored on S3. (3)SQL Editor for deep slicing/dicing of telemetry data with autocomplete, indexed execution, and results in milliseconds. (4)Alerting System that works at the dataset level, supports Slack/Webhook/Alertmanager. (5)RBAC is built in with role creation, user management, and dataset-level access controls. (6)Datasets View showing compression ratios, retention policies, and ingestion statistics.
It's like using a modern IDE for observability, except it runs anywhere, stores on S3, and doesn’t cost you much.
What’s different: (1)Most existing tools force you into their cloud, or assume you’ll glue together five different OSS systems. (2)Parseable + Prism is a single binary, can be hosted anywhere, and gives you all core telemetry modes (logs, metrics, soon traces) in one place.
We’re intentionally minimal on UI fluff, everything is optimized for speed, clarity, and letting engineers move quickly.
Try it out: (1)OSS repo: https://github.com/parseablehq/parseable (2)Live demo (no login required): https://demo.parseable.com/
Related HN posts: Performance Benchmarks Post: https://news.ycombinator.com/item?id=43484310
We would love your feedback, especially from folks maintaining their own observability stacks or running on self-hosted infra. We’re building this from first principles and are happy to answer questions or dig into the details.
Thanks for reading!