The next 3 sections show packet buffer utilization. Resource utilization (%) during last 60 seconds: This section shows session table utilization, which is the maximum number of active sessions supported by the platform. Any value above 80% warrants investigation. Many cores regularly surpassing 75% warrants an investigation.Ġ 42 76 79 85 75 10075 100100100 85 > indicates actual high CPU
This part of the output shows utilization of individual cores on the dataplane. This first section shows CPU utilization by processes running on the dataplane Resource monitoring sampling data (per second): The actual output will look similar to this but may differ slightly depending on the amount of dataplanes and cores your system has. > second Per-second monitoring statistics > minute Per-minute monitoring statistics
If no time operator is used, all views will be listed in one long output. 'minute' shows the last 60 minutes in minute increments and so on. 'second' shows the last 60 seconds of CPU usage in per second increments. Add a time operator to reflect a timeframe you would like to review. 
This command can be used to review dataplane CPU usage. Look for the "-panio" string in the dp-monitor log (this information is logged every 10 minutes) or run the show running resource-monitor command from the CLI to view DP resource usage. When a customer reports a performance issue, generate a tech support file while the issue is occurring. The first step is to isolate where the performance issue is occurring:
If issues are being reported, are these reports coinciding with specific peak times of the day, at regular intervals, or at totally random moments?. Are users experiencing high latency, and if so, does latency affect all traffic or just certain applications?. 
Is some, all, or no traffic through the firewall affected?. Several factors that need to be identified before taking action: Getting an understanding of the impact of the high dataplane CPU is important: high dataplane CPU usage may not present an immediate cause for concern, but being aware of what is causing high CPU and possible consequences will help you respond appropriately. This article shows a couple of ways to verify the actual load of the dataplane and help determine the potential impact on your network. Learn to troubleshoot high dataplane CPU. These factors and more can cause increases in packet rate, packet buffer utilization, or a huge amount of new connections per second that lead to the dataplane CPU reaching a critical level. A number of factors can cause the dataplane's CPUs to spike or continuously run high: a sudden increase due to the implementation of a new service or resource, or a buildup over time due to added connected networks, segments and hosts.
0 kommentar(er)
