While preparing servers for a new service, are you worried that the originally planned hardware will not fit the need in the near future, for how high the load and growth will be cannot be known beforehand? Do you frequently need to expand the network urgently after receiving complaining about unsatisfying efficiency? What is more, unlike suddenly increasing traffic caused by abnormal events, do you find gradually rising traffic and latency hard to detect even though lots of solutions are applied?
Nowadays, sending alerts in advance is more and more important in IT operation. Prediction about low efficiency, increasing traffic, and security events help maintain the performance of network and services.
Data analysis of N-Partner can process and store up to 1 million EPS and can search for data in minimum time. Also, it detects abnormal events and locates the root cause in real time. In the past few years, N-Partner has made IT operation much more effective, benefiting thousands of users. Now, besides storing data for statistics and analyzing to detect anomaly instantly, we take a step forward, making prediction by analyzing received data, providing hardware status and system performance alerts before anomaly happens.
N-Partner’s prediction technology builds a model about future trend with history records. There are currently short-term and long-term prediction, built-in functions in the latest version of both N-Reporter and N-Cloud. As long as the system is under warranty, users can apply them by upgrading the firmware. In short-term prediction, the systems show the trend in the next several hours to few days; as for long-term prediction, there is line graph of the possible amount for months to come. What is more, N-Reporter and N-Cloud will send alerts when more resources are required based on the preset threshold.
Diagram1: The CPU utilization of XX server will reach 95% in 83 days
Diagram2: The interface traffic will be 1 Gbps in 90 days
The built-in prediction function of N-Reporter and N-Cloud is for:
- Hardware: CPU/memory/disk utilization and interface throughput
- Performance monitoring: ICMP round trip time (RTT)
- Website service, including 5 stages of browsing a website: DNS query and responseà TCP connectionà SSLà responseà the first page download
Also, users can enter one or multiple keywords criteria for N-Reporter and N-Cloud to monitor, and the prediction will be the future hit count of events with these keywords.
