Voltage Robotics

Voltage Robotics runs a fleet of edge servers processing sensor data. Before MonSPHERE, "monitoring" meant a cron job that pinged a health endpoint and posted to Slack if it failed — no history, no dashboards, no on-call escalation.

The problem

A three-person platform team doesn't have time to run Prometheus, configure exporters by hand, and wire up Alertmanager routing across a dozen hosts — but they'd outgrown a cron job.

What changed

Voltage scripted the agent installer across their fleet:

for host in $(cat hosts.txt); do
  ssh "$host" "curl -fsSL https://get.monsphere.com/agent | sh -s -- --token=$ENROLL_TOKEN"
done

Each agent connected over mTLS, registered itself, and started supervising node_exporter automatically — no manual Prometheus scrape-config editing across 12 hosts. Voltage then built one dashboard per environment and wired alert rules straight to PagerDuty for anything customer-facing, with a Slack channel for everything else.

Results

Full fleet visibility landed in an afternoon, not a sprint. When a disk-full condition started degrading one edge host two weeks later, the on-call engineer got paged via PagerDuty with the specific host and metric in the alert — the first time Voltage had ever caught a failure before a customer did.

Why it mattered

Voltage didn't hire a dedicated SRE to get here — the same platform a large enterprise runs Platform Admin and custom RBAC on top of also works for a three-person team that just needs real alerting today.