StackStorm

What StackStorm does

StackStorm automates operations work by reacting to events. Sensors, written as Python plugins, watch external systems such as monitoring tools, ticket systems or cloud platforms and fire a trigger when something happens. Rules decide what to do with each trigger: they check its data against criteria and start an action or a workflow, passing the event’s details as inputs. Timers and webhooks are available as generic triggers, so not every automation needs a custom sensor.

Actions are the outbound side. SSH commands and HTTP requests are built in, integrations come in packs, and any script becomes an action once you add a short metadata file. The Orquesta workflow engine chains actions with conditions and shared context, and workflows can be called like any other action. Rules, workflows and actions are all stored as code in packs, so they can be reviewed and versioned like the rest of your infrastructure.

The README describes typical uses: running diagnostics when monitoring reports a failure and posting the results to chat, remediating known problems automatically and calling a human when something goes wrong, and coordinating deployment steps across CI, cloud and load balancers.

ChatOps, approvals and audit

StackStorm was built with ChatOps in mind: teams can run actions and workflows from a chat channel and see the results there. Inquiries pause a workflow until a person answers, which keeps a human in the loop before risky steps. Every execution, manual or automatic, is recorded with its triggering context and results, and the audit trail can be sent to external logging tools. Role-based access control limits who can run what. The project has no paid edition.

Self-hosting StackStorm

StackStorm is made of several services that communicate over a message bus. It runs on 64-bit Ubuntu (20.04 and 22.04) and Red Hat Enterprise Linux or Rocky Linux (8 and 9) only, and the documentation says other distributions are not supported. A one-line script installs the packages on a clean machine. Manual installs, Ansible and Puppet modules, a Docker Compose setup and a beta Kubernetes Helm chart for high availability are also available. The Docker and Kubernetes repositories had not been updated since late 2024 when we checked.

The dependencies are substantial: MongoDB, RabbitMQ, Redis or ZooKeeper for coordination, and nginx in front of the web UI and API. The documentation recommends 2 CPUs, 2 GB of RAM and 10 GB of storage for testing, and a quad-core CPU with more than 16 GB of RAM and 40 GB of storage for production.

Hosted version

There is no hosted service from the project. You run StackStorm on your own servers.

Limitations

StackStorm is an engineer’s tool, not a business automation builder. It has no ready-made connectors for most SaaS apps that office teams use, and building automations means writing YAML, Python or scripts. The footprint is large for a small team, OS support is narrow, and the container deployments get less attention than the core repository, which shipped release 3.9.0 in October 2025.

Who it suits

StackStorm suits operations and SRE teams that want event-driven runbook automation with ChatOps, approvals and an audit trail under an Apache-2.0 license, and that are comfortable running its supporting services. For general data and infrastructure orchestration defined in YAML, compare it with Kestra. For scripts with generated UIs, see Windmill. The workflow automation guide explains where each fits.