I regularly sift through dozens of tech meetup and webinar announcements to filter out marketing fluff and keep only the sessions where engineers share real hands-on experience without abstract slideware. This time, I picked two free infrastructure workshops tackling two of the gnarliest operational headaches: a server suddenly pinned at 100% CPU and elusive network bottlenecks in Kubernetes.
Server at 100% CPU: Isolating the Bottleneck and System Diagnostics
Picture the classic on-call night terror: monitoring alerts are screaming, CPU load is pegged at 100%, the service is timing out, and business wants everything back up yesterday. Your first instinct is to just bounce the process or reboot the VM. Yet if you don't track down the root cause, that CPU spike will be right back in half an hour.
This workshop focuses on staying calm, working step by step, and isolating the fault quickly in a Linux console.
- Organizer: REBRAIN (Rebrain LLC)
- Speaker: Ilya Sachkov, DevOps Engineer at Apteki Plus, author and mentor of the Linux Basics course
- Date & Time: October 23, 2026, at 19:00 MSK (note: rescheduled from October 22)
- Format & Duration: Hands-on online webinar with live Q&A, approx. 90–120 minutes
- Cost: Free
- Who it's for: Sysadmins, DevOps engineers, and backend developers who troubleshoot production incidents
Agenda:
- Choosing your first-response tools: when classic
topis enough (and why hittingPright away to sort by CPU time matters), whenhtopgives better visibility, and when you need deep system telemetry fromatop. - Pinpointing the culprit process: using
pidstatto break load down by thread and tracing syscalls live withstrace -p. - Reading common failure signatures: how syscall patterns and I/O behavior distinguish a memory leak from a tight code loop or an incoming network flood.
Direct registration is open on the event page: Register for the Rebrain webinar.
Kubernetes Networking: Packet Paths, SNAT, and Live Incident Debugging
The second session hits an even trickier boundary—where cloud infrastructure meets Kubernetes virtual networking. Almost anyone who has operated cluster services knows the pain: Kube API suddenly stops responding, a cloud LoadBalancer silently drops client requests, CoreDNS stumbles on resolution, or outbound traffic from pods stalls. Without understanding both the physical and overlay packet paths, you can spend days debugging in the dark.
- Organizer: MWS Cloud Platform (MTS Web Services / MWS LLC)
- Speaker: Roman Erema, Developer on the Managed Kubernetes team at MWS Cloud Platform
- Date & Time: November 3, 2026, at 12:00 MSK
- Format: Online webinar with live demo
- Cost: Free with prior registration
- Who it's for: Kubernetes cluster operators, network engineers, DevOps specialists, and developers deploying services in the cloud
Agenda:
- End-to-end packet path: how the datacenter or VPC physical network, CNI plugin, and internal Kubernetes routing interact from ingress to the target pod.
- Address translation and identity preservation: how SNAT works and why your application logs lose the client's real Source IP.
- Routing pitfalls: when you actually need
externalTrafficPolicy: Local—and why flipping it on can suddenly make your cloud load balancer fail health checks. - Live diagnostic walkthrough: hunting down Kube API, LoadBalancer, and DNS outages on real scenarios in a live demo. All attendees receive a ready-to-use network audit checklist.
You can reserve your spot and join the stream on the official portal: Register for the MWS webinar.

