Anypoint Platform Operations & Administration (Course #platform-ops)

Programme every line is something you do

6 modules · 18 topics · in course-flow order

1. Module 1

   Own the deploy-to-alert loop

   Objective. Standardize deployments, alerts and dashboards in Runtime Manager so daily operations follow a written runbook, not memory.

   • Manage deployments in Runtime Manager: rolling restarts, artifact version pinning and a rehearsed 10-minute rollback drill
   • Route alerts by severity to email and webhook channels (e.g. PagerDuty), separating production from non-production noise
   • Build an operations dashboard per critical application tracking CPU, heap usage and throughput at 1-minute granularity
2. Module 2

   Right runtime per workload

   Objective. Contrast CloudHub and Runtime Fabric administration so every workload lands on the runtime matching its cost and control needs.

   • Weigh CloudHub workers against Runtime Fabric pods: static IPs, inbound bandwidth billing, vCore granularity, upgrade ownership
   • Map workloads: PCI-sensitive flows to RTF inside your VPC, elastic experience APIs to CloudHub multi-worker deployments
   • Run day-2 operations on each platform: CloudHub autoscale triggers, RTF node drain and patching in zero-downtime windows
3. Module 3

   Least privilege by construction

   Objective. Structure users, roles, environments and business groups so every change is attributable and production access stays gated.

   • Assign permissions per persona — deploy-only, read-only, administrator — and split Dev, QA and Production environments
   • Isolate teams through business groups and permission sets; audit membership lists quarterly against the HR directory
   • Review production deploy rights monthly and disable accounts idle for 60 days (target: zero standing admin access)
4. Module 4

   Alert on symptoms, not noise

   Objective. Tune monitoring, retention and thresholds in Anypoint Monitoring so paging fires only for user-visible degradation.

   • Monitor golden signals — p95 latency, error rate, saturation — with threshold hysteresis (e.g. breach held 5 minutes)
   • Split log retention: 7 days hot in Anypoint Monitoring, 90 days searchable in the central SIEM via forwarding
   • Suppress flapping alerts with deduplication and quiet hours; target fewer than 5 actionable pages per on-call week
5. Module 5

   Certificates that never page you

   Objective. Automate keystore inventory and renewal so every certificate rotates ahead of expiry without an emergency change window.

   • Inventory keystores and certificates across environments with expiry alerts firing 30 days out
   • Install CA-signed TLS chains with full SAN coverage on CloudHub and RTF listeners; ban self-signed certs from production
   • Rotate without downtime: staged keystore upload, blue-green listener switch, then an OpenSSL s_client handshake check
6. Module 6

   Rehearse the Disaster You Fear

   Objective. Prove Vendora's RTO < 15 min and RPO 0 claims in a drilled failover, not in a slide deck.

   • Script the region-failover runbook — DNS cutover, worker redeployment, Anypoint MQ drain — timed against RTO < 15 min
   • Verify RPO 0: persistent Object Store and transactional messaging survive a killed worker; restore and reconcile state
   • Run semi-annual game days: inject zone loss, grade the response, feed every gap into the platform hardening backlog

References & Official Documentation every claim above has a source

Module 1

• Runtime Manager AlertsMuleSoft Docs
• Using Built-in Mule App DashboardsMuleSoft Docs

Module 2

• Anypoint Runtime Fabric OverviewMuleSoft Docs
• CloudHub 2.0 ArchitectureMuleSoft Docs

Module 3

• Permissions in Anypoint PlatformMuleSoft Docs

Module 4

• Getting Alerts in Anypoint MonitoringMuleSoft Docs

Module 5

• Configure TLS with Keystores and TruststoresMuleSoft Docs

Module 6

• CloudHub 2.0 High Availability and Disaster RecoveryMuleSoft Docs

More in Operations / DevOps