Getting started
Getting started with infrastructure automation
How to choose your first automation, scope it so it cannot do damage, and prove it works before you expand.
9 min readRead guide
AI-powered infrastructure automation
Infrastructure Automation, Without the Operational Complexity.
From support tickets to cloud provisioning and backup operations, StackHeal helps you automate the work behind modern infrastructure — so your team can move faster, deliver better service, and scale with confidence.
Every automated action is scoped, approvable and verified — see how the controls work.
Automation runrun_4821
01Request
Ticket #10482
“Disk usage warning on web-07 — site is slowing down.”
- Source
- Service desk
- Asset
- web-07 · AMS-1
- Intent
- Capacity issue
02Approved workflow
Reclaim disk space
wf.disk_reclaim.v3
- Gather context and metrics
- Policy and permission check
- Rotate logs, purge temp cache
Scoped to approved actions
03Verified outcome
61%
94%
Disk usage after run · 42 GB reclaimed
- Result re-checked against threshold
- Ticket updated and closed
- No engineer touched a console
Connects to the tools you run
- Proxmox VE
- VMware vSphere
- Zabbix
- Prometheus
- WHMCS
- Jira Service Management
- Ansible
- REST + webhooks
Product ecosystem
Everything You Need to Automate Infrastructure Operations.
Bring the workflows that run your infrastructure into one connected automation platform.
01
AI Support Automation
Turn incoming support requests into structured, actionable workflows.
Learn more02
Cloud Provisioning
Standardize virtual machine provisioning and infrastructure lifecycle operations.
Learn more03
Backup Automation
Coordinate backup jobs, monitor outcomes, and simplify recovery procedures.
Learn more04
Infrastructure Operations
Connect monitoring signals with troubleshooting and approved remediation workflows.
Learn more05
Workflow Orchestration
Connect systems and automate repetitive operational processes.
Learn more06
Integrations
Bring support, monitoring, and infrastructure systems together.
Learn moreAlso available: Server and Data Center Operations — scheduled, staged operations across hosts, racks and sites.
The platform
One Platform. Connected Operations.
Support requests, infrastructure resources, monitoring signals, automation workflows and verification all meet in one place — instead of being stitched together by people moving between consoles.
StackHeal
Connects every system below
- Customer supportTickets and requests
- Infrastructure resourcesHypervisors, clouds, servers
- Monitoring and alertsMetrics, alerts, incidents
- Automation workflowsPlaybooks and approvals
- Verification and visibilityOutcome checks and audit trail
Fewer disconnected processes
The handoffs between support, infrastructure and monitoring stop being manual relays and become steps in a single run.
Better visibility
One run history answers what happened, on which asset, under whose approval, and whether the outcome was verified.
More consistent execution
The same workflow definition runs the same way across shifts, sites, tenants and platforms.
Solutions
Built for the Businesses That Keep the Digital World Running.
Different operations, same underlying problem: too much of the work between a request and a result is done by hand.
- Data CentersSimplify recurring operational tasks, improve workflow consistency, and reduce manual coordination.
- Cloud and GPU Cloud ProvidersStandardize provisioning requests, infrastructure workflows, and support operations.
- Hosting ProvidersStreamline customer requests, routine server operations, and service delivery.
- Managed Service ProvidersAutomate repetitive tasks across customers, systems, and operational workflows.
- Enterprise IT TeamsConnect service management with controlled infrastructure automation.
How it works
From Request to Resolution. With Less Manual Work.
One real workflow, stage by stage. Each step is observable, and each step can be configured to stop and ask a person.
Step 1 of 5
A request arrives
A customer or colleague submits a ticket describing a symptom, in their own words. Nothing about the request is structured yet.
What exists at this stage
- Ticket
- #10482
- Body
- “Disk usage warning on web-07 — site is slowing down.”
- Source
- Service desk webhook
Illustrative values from a sample run.
AI in operations
Put AI to Work on Real Infrastructure Challenges.
StackHeal uses AI where interpretation genuinely helps — reading an operational request, understanding its context, assisting a diagnosis, matching a runbook. What it does not do is decide what is permitted.
Interpretation proposes. Policy permits.
Permission, scope and blast radius are evaluated in code against a registered action catalogue — not inferred by a model. Sensitive operations can require an approval before anything runs.
01
Support requests that trigger relevant workflows
A request written in prose is classified into your own operational taxonomy, resolved to a specific asset, and matched to the workflow you have approved for it.
Interpretation proposes. Policy permits.
02
Infrastructure alerts that start a guided investigation
Related signals are correlated into one incident, the read-only checks a first responder would run are executed, and the pattern is matched against your runbooks and past incidents.
Diagnostics are read-only by default.
03
Provisioning requests that follow standard policy
An order or ticket is normalised into a specification, checked against entitlement and quota rules, and built from an approved template with verification afterwards.
Only registered templates and sizes are selectable.
Registered actions only
Workflows compose from an action catalogue you control. There is no general-purpose shell for a model to reach for, and it cannot introduce a new capability at run time.
Permission decided in code
Whether an action may run on an asset is a deterministic policy evaluation, not model output. Scope, tenant and environment are resolved outside the model's influence.
Approvals where risk warrants them
Sensitive operations pause for a named approver, who sees the resolved asset set and the exact parameters before deciding.
Verified, or escalated
Outcomes are asserted against measured state. A run that cannot verify itself escalates with its findings instead of reporting success.
Why StackHeal
Simpler Operations. More Room to Grow.
No invented savings percentages. These are the four things teams consistently report after the first few workflows are running.
Reduce manual work
Spend less time on the repetitive operational tasks that make up most of a queue — and keep your engineers on the problems that actually need them.
Deliver faster
Standardise the path from request to completion so work moves at the same pace regardless of who is on shift or how deep the queue is.
Improve visibility
See what ran, what it touched, who approved it, whether it was verified, and what failed — from one run history rather than four consoles.
Scale confidently
Reuse the same workflow definitions as you add customers, sites and platforms, with per-tenant policy instead of per-tenant procedures.
Integrations
Works With the Tools Behind Your Infrastructure.
StackHeal is the layer between the systems that generate operational work and the systems that do it. Both sides connect through maintained connectors or through APIs and webhooks.
Ticketing and service management
Where requests arrive and where outcomes need to be written back.
- WHMCS
- Jira Service Management
- Zendesk
- Freshdesk
- +2 more
3 native connectors · 6 total
Virtualization and cloud platforms
The platforms that actually run the workloads.
- Proxmox VE
- VMware vSphere
- KVM / libvirt
- Microsoft Hyper-V
- +2 more
2 native connectors · 6 total
Monitoring and alerting
The signals that start an investigation.
- Zabbix
- Prometheus Alertmanager
- Grafana Alerting
- Nagios-family checks
- +1 more
2 native connectors · 5 total
Backup and storage
Systems that hold recovery points and archive data.
- Proxmox Backup Server
- Veeam
- Restic and Borg
- S3-compatible object storage
- +1 more
3 native connectors · 5 total
Automation and scripting
How workflow steps are executed on target systems.
- Ansible
- Shell and PowerShell
- Terraform
- Redfish and IPMI
2 native connectors · 4 total
APIs and webhooks
The generic path for anything without a dedicated connector.
- StackHeal REST API
- Outbound webhooks
- Generic HTTP action
- Chat notifications
4 native connectors · 4 total
On connector status: the directory labels every system as a native connector, a beta connector, API-based, or planned. We do not imply a working native connector exists for every product named — check the label, and ask us if a specific one matters to your plan.
Resources
Learn How Teams Automate Operations.
Practical guides written for the people doing the work — how to scope a first automation, structure provisioning, make backups verifiable, and use AI in operations without handing it the keys.
Automating cloud provisioning workflows
A reference structure for provisioning automation: intake, policy, validation, build and post-deployment verification.
11 min read
Building reliable backup operations
Reconciliation, retry policy, retention and restore testing — the four things that separate backup coverage from backup confidence.
10 min read
Connecting support tickets to infrastructure actions
How to close the loop between a request system and the infrastructure behind it without handing automation unbounded access.
10 min read
Best practices for AI-assisted IT operations
Where language models genuinely help in operations work, where they should not be trusted, and the controls that make the difference.
11 min read
Ready to Simplify Your Infrastructure Operations?
Discover how StackHeal can help your team connect systems, automate repetitive work, and deliver infrastructure services more efficiently.