Live system monitoring and I/O profiling at scale
Mistral is a lightweight profiling tool for full system telemetry or wide scale profiling.
Protect shared storage from rogue jobs and noisy neighbour applications by running Mistral in production.
Run as part of a QA or continuous integration framework to find bad I/O patterns
Case studies
Storage protection at ARM
“The data and system control provided by Mistral allows the infrastructure teams to prevent risky I/O patterns and give us a lot more information to learn from.”
Catching bad I/O patterns
“We really like Mistral – it’s really smart. We have nearly eliminated rogue jobs on the cluster.”
Profiling containers
“Singularity containers for high performance and AI, coupled with the tools from Ellexus, provide critical insights into application life cycle and tuning.”
Take control of the way applications access data
Mistral can be deployed in production, in test or as a one-off system healthcheck. By following our best practices you can ensure that your applications and infrastructure are fast, agile and cloud ready.
Collect per-job read, write and meta data metrics
Set different policies for different file systems
Analyse system trends for specific jobs, users, groups or projects
Integrated with Grafana, InfluxDB, Elastic search, Splunk and more for live system overview
Throttle problem applications by setting IOP and bandwidth limits
Benchmark in production for procurement and scaling
How much time are you spending in good and bad I/O?
- Run a system wide I/O healthcheck to find out how much time is spent in good and bad I/O
- System-wide dashboard for live overview: Integrate with existing analytics framework easily
- Deploy Mistral in your QA test environment to prevent bad I/O from reaching production.
- Allow users to opt in to more detailed I/O profiling to check performance and application scalability.
- Customisable reports based on individual applications, users, groups or projects.