Operate

Know what is running, what changed, and why it broke.

A team with nobody on call still needs answers on day two. SysJ keeps an inventory of every environment, notices when reality stops matching the plan, explains failed deployments, and tears things down safely when you are done.
drift · exampleNeeds a decision
Resourcesg-api-ingress
Declared443 from load balancer
Found443 from 0.0.0.0/0
Changed inAWS console
Optionsadopt · revert

At a glance

What Operate gives you.

One inventory

Every cluster, database and bucket SysJ manages, grouped by environment, with what it is for and when it last changed.

Drift, surfaced

A change made in a cloud console is detected and shown to you. SysJ never silently reverts a fix someone made during an incident.

Failures explained

When a deployment fails, SysJ shows the step that failed and the events around it, in plain language, before you open a terminal.

Teardown that shows its work

Destroy is a plan like any other. You see every resource that will go, with anything holding data listed separately.

In detail

How it works.

Drift

Quiet until it matters, then specific.

SysJ compares declared state with your account on a schedule. Changes that carry no risk stay quiet. A change that opens a port or removes encryption becomes a notification that names the resource, the difference, and who can decide.

  • Adopt the change into the declaration, or revert it
  • Either choice is a plan, approved like any other
  • Resources SysJ created and no longer declares are flagged
drift · exampleNeeds a decision
Resourcesg-api-ingress
Declared443 from load balancer
Found443 from 0.0.0.0/0
Changed inAWS console
Optionsadopt · revert

Teardown

Hard to do by accident, easy to do on purpose.

Removing an environment produces a destroy plan that names every resource in your language rather than in HCL. Databases and volumes are called out on their own with their own acknowledgement. Production is protected by default.

  • Deletes in dependency order
  • Protected resources cannot be removed by clicking through
  • Resumes cleanly if a teardown stops partway
destroy plan · exampleHolds data
Clusterstaging · 2 nodes
Networking-8 resources
Databaseorders-staging · snapshot taken
Acknowledge data lossrequired
Confirm by typingstaging

Health and history

Is it up, and what happened.

Each environment shows the health of its cluster and workloads, the last deployment, and a timeline of every change and approval. Notifications go to Slack or PagerDuty when something needs a person.

  • Cluster and workload health per environment
  • An append-only activity trail for the whole workspace
  • Slack, PagerDuty and webhook notifications
production · example
Clusterhealthy · 3 of 3 nodes ready
orders-api4 of 4 pods ready
payments-workerrestarting · OOMKilled
Last deploymentv1.14.0 · 2 hours ago
Driftnone

Questions

Common questions about Operate.

What happens if someone changes a resource in the AWS console?
SysJ detects the difference on its next comparison and shows it to you. You choose whether to adopt the change into your declared state or revert it. Nothing is reverted automatically.
Can SysJ take over infrastructure we already have?
Yes. Import reads an existing account into declared state and produces a plan that is empty where reality already matches, which proves the model is correct before anything changes.
Is production protected from being destroyed?
Yes, by default. Removing a protected resource requires editing its declaration first, which is reviewed and attributed like any other change.

See your own infrastructure as a plan.

Tell us about your team and the cloud you use. We will set up your workspace and walk through your first plan with you.