Job lifecycle management
Submit jobs from uploaded JARs, resume from savepoints for seamless upgrades or recovery, and safely stop or cancel them when needed.
Problem
Your Flink footprint has grown, your visibility into it hasn't.
Even when jobs are running, your team is often still:
- Digging through logs and the Flink web UI just to understand why a job is backpressured
- Routing every job submission, savepoint, or cancellation through the one engineer who knows the cluster
- Piecing together metrics from Prometheus, thread dumps, and task manager logs to debug a stuck task
- Managing a fleet of clusters spread across environments with no single place to see them all
- Trying to enforce access control and audit requirements without native RBAC or SSO in Flink itself
Solution
A single control plane for Apache Flink in production.
See everything
Understand your Flink jobs at a glance, with the complete picture from job graphs to checkpoint history.
- Real-time visibility into jobs, checkpoints, metrics, configs, and state
- Job graphs with live metrics at every operator, plus exception and backpressure analysis
- Unified, real-time view of cluster health, memory and CPU usage, task slots, and throughput
- REST API and Prometheus integration for your existing monitoring and alerting stack
Debug and troubleshoot
Drill into individual jobs and tasks to find and fix bottlenecks without stitching together logs by hand.
- Visual inspection of real-time job graphs, exceptions, and operator backpressure and watermarks
- Live task manager logs, thread dumps, and memory monitoring for stuck or underperforming tasks
- Checkpoint and savepoint monitoring to validate pipeline behaviour end to end
- Task-level metrics to quickly identify which operator is the source of a problem
Act with confidence
Manage the full operational lifecycle of Flink jobs from a clean UI backed by a REST API.
- Job submission from uploaded JARs, resume from savepoints, and safe stop or cancel
- Savepoints for seamless upgrades or state backups without disrupting running pipelines
- A single control plane across co-located or hybrid and cloud Flink clusters
- Works with open-source Apache Flink and Ververica Platform from one interface
Stay secure
Enterprise-grade governance for teams that need more than what Flink ships with natively.
- Role-based access control, SSO, and multi-tenancy across your entire Flink fleet
- Data masking to redact sensitive data for every user and every action
- Complete audit trail of every action taken in the cluster
- Full support for air-gapped deployments with zero data egress
Outcomes
How teams use Flex
Deep job debugging
Visually inspect real-time job graphs, analyse exceptions, and drill into operator backpressure and watermarks to fix bottlenecks.
Live task analysis
Inspect individual task managers, view live logs, analyse thread dumps, and monitor memory to troubleshoot stuck or underperforming tasks.
Platform observability
Get a unified, real-time view of your cluster's health, including memory and CPU usage, task slots, and data throughput across all jobs.
Stream and state QA
Manage and monitor checkpoints and savepoints, and validate pipeline behaviour by inspecting task-level metrics, logs, and watermarks.
Governance and security
Apply fine-grained access rules, redact sensitive data, and maintain audit trails for every action taken in the cluster.
Deployment
A Flink control plane that fits the way your team already deploys.
Flex works seamlessly with a growing number of Flink distributions, including open-source Apache Flink and Ververica Platform, so you keep one consistent control plane across your entire Flink ecosystem.
Runs on every Flink distribution
Apache FlinkVerverica PlatformApache FlinkVerverica Platform
Deployment
No new infrastructure, running in minutes.
Immediate value
Go from installation to full visibility in minutes. A streamlined, lightweight setup means you connect to clusters and manage jobs immediately, skipping complex configuration.
Deploys anywhere
Runs as a single Docker container or a single JAR file, on-premises, on Kubernetes via Helm charts, or on ECS and Fargate.
Air-gap capable
Full support for air-gapped deployments with zero data egress, a low memory footprint, and low CPU utilisation.
Per-cluster pricing
Flex is priced per cluster rather than per user or per seat, so your costs stay flat as your engineering team grows.
Customers
Results from teams using Flex in production
Flex is proven in mission-critical environments, and for Kpow users it provides a seamless, unified experience across your entire data-in-motion stack.
“I am grateful for the empathy and passion the Factor House team has shown in partnering with Airwallex.”
Alex Hilton
Chief of Staff-Technology, Airwallex
Pricing
Start free. Scale without surprises.
Priced per cluster, not per seat. Your costs stay flat as your team grows.
Community Edition
Free Forever · No credit card required
For individual developers and learning.
Start free- Essential Flink job monitoring
- Visibility into up to 3 clusters (non-production)
- Community Slack and docs
Enterprise
$3,950 / cluster / year
No per-seat fees.
Speak to the team- SSO and role-based access control
- REST API, Prometheus, and webhook integration
- Administrative workflows and audit logs
- Unlimited clusters
- Priority email support
Experience better Flink operations first-hand
Start with the free version and have Flex running in your environment in minutes via Docker, Kubernetes, or JAR. If you're evaluating Flex at scale, get in touch and we'll walk through how it fits your environment.