Operate your infrastructure, not just monitor it.
Ecropolis ControlPlane brings infrastructure health, environment readiness, cache operations, release workflows, and platform diagnostics into one guarded operating surface. Built for teams running high-stakes web platforms across multiple nodes, caches, load balancers, and edge layers.
Operations break down when the site becomes a fleet.
Most tools manage one install, one update, or one server. A serious platform doesn't run that way — it runs as multiple delivery nodes behind a load balancer, an origin or author node, layered caches, CDN rules, and a deployment pipeline.
Incidents rarely come from one obvious failure. They come from drift, stale caches, overloaded nodes, and unclear traffic state — the "is delivery-3 actually the problem?" question. ControlPlane gives operators fleet-level context before they touch production.
One operating surface for the whole fleet.
Fleet & Node Control
- Live topology model of every node — delivery, author, cache, and search
- Node lifecycle stages with the next required action always visible
- Load balancer backend management: attach, drain, accept, weight
- Config drift detection that catches the node that silently differs
Cache Verification
- Coordinated invalidation across page cache, object cache, and CDN edge
- Verified purges — done means proven miss→hit re-render, not "API said OK"
- Stale-URL diagnosis that reports exactly which layer is lying
- Durable purge queue with depth visibility, priority, and a kill switch
Health & Observability
- Per-node readiness, render timing, and PHP worker signals
- 5xx, fatal, and deadlock visibility from live log analytics
- Redis and database health, windowed — not lifetime averages
- Node peek: fetch any URL from a specific node and inspect the result
Guarded Production Actions
- Recommendation-first: the console proposes, an operator approves
- Manual approvals, holds, and required change reasons
- Full audit history for every meaningful action
- Slack notifications with the reasoning attached
WordPress Operations
- Staged core and plugin updates: stage → verify → flip → rollback
- Plugin lifecycle jobs across the whole fleet
- Environment-scoped feature flags with rollout ladders
- Rollback-aware workflows for every change
AI-Assisted Operations
- AI-generated findings, evidence, and next checks from live signals
- MCP tools so AI agents can diagnose with the same guardrails as humans
- AI can investigate; production actions stay human-approved
Built for actions with receipts.
Dashboards tell you what a system claims. ControlPlane verifies outcomes — cache purges, node readiness, balancer state, release gates — and logs every meaningful action with its reason, result, and evidence.
- API success is not proof. Verification is proof.
- Every action has a reason, result, and audit trail.
- AI can investigate, but guardrails stay in place.
The workflows operators actually run.
Diagnose a stale page
Probe a URL through every cache layer on every node and see exactly which layer is serving stale content.
Drain and validate a node
Take a delivery node out of rotation, verify its health, and return it to traffic — without guesswork.
Rotate fleet reboots
Plan and rotate OS reboots across the fleet without dropping capacity.
Stage and roll back updates
Stage WordPress core and plugin updates, verify them, flip — and roll back cleanly if needed.
Catch config drift early
Per-node config snapshots surface the machine that differs from its siblings before it becomes an incident.
Escalate edge protection
Arm CDN challenge modes during sustained distress — without mistaking your own purge traffic for an attack.
Your infrastructure, one control plane.
ControlPlane sits alongside your stack — reading health and log signals from every layer, and acting on them through guarded, audited workflows.
Software plus the operators who know how to use it.
ControlPlane is delivered and operated with Ecropolis. We can provide the software, run your platform end to end, or meet you in between — the same team that builds the console operates fleets with it every day.
Explore our support & maintenance services →How this work actually goes.
Engineering write-ups from real fleets — the problems that led us to build ControlPlane, and what it took to fix them.
Case Study: A Million Broken URLs
What happens to a 20-year-old website's links when the platform changes — and the bots that never stop knocking.
Read the write-up → EngineeringCase Study: Faster and Cheaper
Most capacity conversations end with a bigger bill. This one didn't.
Read the write-up → EngineeringCase Study: Four Servers, Four Snowflakes
Hand-managed fleets don't fail loudly. They drift quietly — until every server is subtly, differently wrong.
Read the write-up → EngineeringCase Study: The Two-Minute Promise
The hardest problem in web performance isn't making a site fast. It's making a fast site update.
Read the write-up →See ControlPlane on your fleet.
Tell us how your environment is set up. We'll show where ControlPlane fits, what it can verify, and which operational risks it can reduce first.