Computation Performance
Monitoring
Cut costs or increase performance up to x5 times by leveraging cutting edge AI-engine taking advantage of careful performance monitoring
2 minutes integration
Simple script to install data collection agent or integrate with your
monitoring tools
10k runs in database
We leverage more than >100GB of performance metrics to give you suggestion after first day of monitoring
Secure
No hidden logic.
Data collection agent is open source. We see only what you allow
powered by Typeform
powered by Typeform
Pricing
FREE
  • Performance stats
  • Grafana visualization
  • CPU/RAM size suggestions for your workloads
$0
Join now
Single
  • Advanced CPU/RAM/Storage stats
  • Precise suggestion to exact cloud hardware shape on AWS, GCP or OCI
  • Software suggestion: OS version, drivers, solvers, compilers
$49/month
Join now
Enterprise
  • Network and openMPI stats
  • Cluster configuration suggestion with RDMA/EFA/gluster set up
  • Datadog and Newrelic integration.
$1449/month
Join now
Plans to fit your scale
Features
FREE
Single
Enterprise
Performance stats
Grafana visualization
CPU/RAM size suggestions for your workloads
AWS shape config suggest
Google GCP shape suggest
Oracle OCI shape suggest
Software suggestion: OS version, drivers, solvers, compilers
Network and openMPI stats
Cluster config set up with RDMA/EFA/gluster
Datadog and Newrelic integration.
How it works
1
Create tracking id
Specify operation system and contact email
2
Copy
Copy and past few scripts commands. Analytics set up is done.
3
Run your workload
You will get notification that collection is started
4
Log in to performance monitor panel
We sent credentials to your email
5
Help ML Algo with workloads grouping
Every run add its own fingerprint. Performance matching ML Algorithm trying to match similar workloads. You can help algo with manual match
6
Check out configuration siggestions
After 1-7 launches algo will suggest you (faster or cheaper) configuration on Amazon AWS, Google GC or Oracle OCI. You can launch configuration right from our control panel or copy info.
CPU
  • input output waiting >10% for 5 minutes
  • >5% virtual CPU waits for real CPU for 5 minutes
  • >5% interrupt requests for 5 minutes
  • >80% usage for 5 minute
DISK
  • >70% of disk.total for more than 5 minutes
  • >50% weighted input output time for more than 5 minutes
Memory
  • <25% available memory :for more than 5 minutes
  • <50% portion of a storage that is used for virtual memory
  • >5% of dirty memory for more than 5 minutes