LoginSign Up
← Back to all sparks
P

Porter

INFRA · APIS
Velocity0.0

Platform-as-a-service that runs and scales applications inside your own AWS, GCP, or Azure cloud account

porter.run ↗

Porter tightens its security layer with short-lived cloud credentials and zero-trust workload access.

kubernetespaascloud-infrastructuresecuritygpudeveloper-tooling
◆Current state
Porter is a Kubernetes-based PaaS targeting engineering teams that want AWS/GCP infrastructure without managing raw k8s. Over the past year it has added GPU job support, temporal-based autoscaling, multiple node groups, and overhauled both application creation and datastore management flows. The most recent release focuses squarely on access control: short-lived cloud credentials and secure cloud access for workloads indicate a move toward zero-trust IAM patterns at the infrastructure level.
◆Where it's heading
Porter is building a security and compliance layer on top of its deployment primitives — audit logs appear twice across these entries, SOC 2 controls via Drata, short-lived credentials, and workload identity. In parallel, AI-adjacent features have appeared at each release cycle: DeepSeek-R1 model deployment, a Helicone observability add-on, GPU time-slicing, and full GPU Jobs. The two tracks — security hardening and AI infrastructure — are converging toward a production-ready positioning for AI workloads.
◆Prediction
The next release will likely push deeper on IAM: workload-level RBAC and service account policies are the logical follow-on to short-lived credentials. Expanded GPU options for training (not just inference) jobs are probable given the incremental progression from time-slicing to full GPU Jobs.

◆Recent moves

  1. 2mo ago

    API Audit Logs, CLI and Browser Session Management, Short-lived Cloud Credentials, and Secure Cloud Access for Workloads

    ⚡ SPARK

    Short-lived cloud credentials and secure workload access represent a meaningful architectural shift — Porter is moving from long-lived static credentials toward ephemeral, workload-scoped IAM patterns that reduce the blast radius of a compromised environment. This is the clearest signal yet of a security-first positioning.

  2. 8mo ago

    New Application Create Flow, Revamped Datastore Create and Update Flow, Multiple Node Groups, and Temporal-based Autoscaling

    Temporal-based autoscaling and multiple node groups are substantive infrastructure additions — the first enables time-aware scaling (predictable traffic patterns), the second lets teams mix instance types within a cluster. The revamped datastore and app creation flows reduce the configuration surface for day-one setup.

  3. 8mo ago

    Audit Logs, Custom Autoscaling, Persistence for the Grafana Add-on, GPU Jobs, and Improved Deployments

    GPU Jobs and custom autoscaling are the headline additions here — GPU Jobs allows batch GPU workloads without a persistent running service, which matters for ML training tasks. Persistence for the Grafana add-on and improved deployments address gaps in the observability and rollout experience.

  4. 1y ago

    Datastore Updates, Metrics and Logs v2, Dismiss Notifications, Cancel Pre-deploy Jobs

    Metrics and Logs v2 is a meaningful revision to Porter's observability layer — a v2 designation typically signals a rethought data model rather than cosmetic changes. Datastore updates and the ability to cancel pre-deploy jobs round out operational control improvements.

  5. 1y ago

    Tailscale Integration, Cost Optimization Granularity, Helicone Add-on, and Improved App Alerts

    Tailscale integration enables private networking between workloads without opening public ingress — a security-conscious addition that fits the zero-trust direction Porter is building. The Helicone add-on for AI inference observability continues the pattern of bundling AI-adjacent tools into the platform.

  6. 1y ago

    Multi-app porter.yaml, Improved CLI and Audit Logs, Faster Docker Build Times

    Multi-app porter.yaml brings monorepo teams and multi-service architectures into a single declarative config file — a quality-of-life improvement that reduces context-switching between app configurations. Faster Docker build times are a welcome operational gain with no behavior change required.