Intermediate everything in DevOps & Cloud
Datadog Infrastructure Basic targets host health checks
Datadog describes a lighter monitoring tier for host metrics and selected processes on noncritical workloads.
CNCF argues agents should propose system state, not receive production root
A CNCF Ambassador argues that agents should express changes as versioned system state and pass them through review, build and deployment controls instead of receiving unrestricted production root.
CNCF announces its 2026–2028 Ambassador cohort
CNCF accepted 58 new ambassadors for its 2026–2028 cohort, while 48 previous ambassadors stepped away.
ArgoCon previews the direction of Argo CD 4.0
CNCF previews ArgoCon North America 2026 and says the Argo community is beginning the visioning process for Argo CD 4.0. The event covers four distinct projects—Argo CD, Workflows, Rollouts, and Events—so learners can map each project to a delivery problem before treating the vision as a release.
AWS introduces CloudWatch Omni for collaborative observability
AWS announced CloudWatch Omni on 22 September, an OpenTelemetry-based experience for application and AI-agent observability outside the AWS console.
Kubernetes 1.37 adds a native unused-PVC condition
Kubernetes 1.37 promotes PVC unused tracking to Beta, adding an Unused condition managed by the PVC protection controller.
Docker publishes Sandbox Kit Specification v3
Docker published an open-source Kit Specification v3 for packaging an agent’s runtime, tools, permissions, and network policy with its environment.
GitHub describes an agentic fuzzing taskflow
GitHub Security Lab describes a taskflow that finds entry points, writes harnesses, runs AFL++, and triages crashes for C/C++ projects.
AWS adds shared, ordered EventBridge custom buses
AWS announced enhanced EventBridge custom event buses that can be shared across accounts, support ordered delivery, and use a Subscriber resource.
Grafana 13.2 adds multiple notification policy trees
Grafana says multiple notification policies are generally available in Grafana-managed alerting with Grafana 13.2, enabling independently managed routing trees.
GitLab publishes critical patch releases 19.4.1, 19.3.3 and 19.2.7
GitLab released 19.4.1, 19.3.3 and 19.2.7 on 23 September with important bug and security fixes for self-managed installations.
Docker introduces Cloud Sandboxes for long-running agents
Docker announced Cloud Sandboxes on 24 September: the same microVM isolation model as local Sandboxes, running on Docker-managed compute.
Critical vm2 sandbox escape (CVSS 10.0) traced to its own README example
GitLab's Threat Research Group found a CVSS 3.1: 10.0 sandbox escape in vm2, a widely used Node.js sandbox library, in a config copied from vm2's own README. Version 3.11.7 fixes the attack, but the underlying risk remains for anyone using require.external. Update immediately and prefer containers for untrusted code.
Federating reliability ownership to product teams
Build a self-service reliability platform (golden paths, paved roads), train teams and embed SLO/on-call practices, and govern with standards plus error budget…
Capacity planning for a distributed cache
Track hit ratio, memory and eviction rate, throughput/latency, and connections; correlate with growth to forecast when to add capacity before the hit ratio or evictions…
Capacity planning for distributed stateful systems
Account for replication and cross-region network, IOPS and disk throughput, connection limits, partition/shard balance, and N+1 region failover headroom; validate with load and chaos…
Designing a feature flagging service
Control plane (UI, store, targeting), SDKs that cache flags locally for zero-latency evaluation, streaming/CDN delivery for near-real-time updates, and stale-flag lifecycle tooling.
Managing a risky release with feature flags
Deploy code dark behind an off flag, enable for internal then small percentage, monitor metrics, ramp gradually, then remove the flag.
Automating actions on error budget burn
Use multi-window burn-rate alerts; on fast burn, auto-trigger actions like halting deploys, rolling back, or scaling, with escalating tiers.
Safeguards for automated remediation runbooks
Add rate limits and circuit breakers on restarts, escalate to humans after N attempts, log all actions, and check for cascading failure before acting.
We are hiring for this. Every open role lists the topics its interview covers, so you can prepare for the real thing rather than guessing.
See open roles