I’m working on OpsOrch(https://www.opsorch.com/), an open-source orchestration layer that provides a single API across incidents, logs, metrics, tickets, messaging, and service metadata.

It sits on top of existing tools like PagerDuty, Jira, Prometheus, Elasticsearch, and Slack, and normalizes them into a shared schema. It doesn’t store operational data, it just brokers requests through pluggable adapters and returns unified structures.

The motivation came from incident response workflows that still require hopping across multiple vendor UIs and APIs with different auth models and query languages. Instead of another “single pane of glass,” this is meant to be a small, transparent glue layer.

On top of the core service, I’m also exposing everything via an MCP server so LLM agents can query incidents, metrics, and logs as typed tools without needing vendor-specific knowledge.

Currently open source, written mostly in Go and TypeScript. Still early, but usable with PagerDuty, Jira, Prometheus, Elasticsearch, Slack, and mock providers. Feedback from SREs and infra folks has been very helpful so far.