Announcing Flowmill's Integration with Nomad
Nomad has always stood out for its flexibility and breadth of support for real-world environments. The migration to containers is rarely an all or nothing proposition and frequently SRE teams find themselves managing hybrid environments spread across on-premise and public clouds. Getting visibility into these complex systems is incredibly important.
At Flowmill, we’re razor-focused on improving observability in these types of complex environments and we are excited to be working with the Nomad community. The integration available today allows Nomad users and administrators to monitor containerized and non-containerized workloads using the Flowmill platform.
How it works
Flowmill tracks network behavior from within the operating system and connects it with metadata from an orchestration platform. In a Nomad environment, Flowmill automatically captures information about jobs, groups, and tasks running on each node and merges it in real-time with network socket data gathered from the operating system. This provides users a unified, contextual view of workloads distributed throughout the cluster without any code changes or tagging.
With this approach, Flowmill can help Nomad users in a number of ways including:
- Analyzing traffic flows in a cluster to discover dependencies or identify changes in behavior
- Measuring network transfer costs and latency resulting from cross-zone or cross datacenter traffic
- Analyze traffic to identify network issues, including increases in latency and connection failures
All of the above can be viewed for entire jobs or users can drill down to see the behavior of specific tasks.
Rob Cameron, Technical Director at Roblox, noted:
Orchestrators do their best at packing as many jobs as possible into hosts. This is amazing from a compute perspective, but it can strain the network of a device. It can become impossible to determine how applications interact with each other on a network level. This can lead to unexpected results for applications leaving teams scratching their heads to figure out what is going on. Using Flowmill I can easily determine how applications communicate not only on a device level, but across my entire infrastructure. This deep knowledge can resolve finger pointing with real-time network tracing. The integration with Nomad allows me to correlate jobs, groups, and tasks across thousands of hosts with very little effort on my part.
The Flowmill team found the integration with Nomad to be really easy given its straightforward APIs and fantastic documentation. We’re excited to work with more Nomad users to help them improve observability! If you are interested in seeing this in action, please reach out for a live demo or check out Youtube channel.