The Calendar Heat Chart views help engineers and managers identify patterns in utilization, latency, and loss metrics. Identifying such patterns can be a critical step in locating the source of performance issues that appear to be intermittent. Usage patterns are also useful for capacity planning for circuits, routers, servers, and other devices. Use Calendar Heat Charts to configure time-based access lists or QoS policies.
Four colors are used as “heat” indicators to rate the performance of the selected metric for each individual hour of the day. The default values for these color ratings are based on industry best practices. You can customize the heat indicator values in the View Settings dialog. Click the gear icon on the view to open the dialog.
Across the top of each Calendar Heat Chart, a standard abbreviation (such as “Mo” for Monday) indicates the day of the week. Just below the day of the week, two rows report the month and date. With this information, you can correlate the data in the Calendar Heat Chart by date or by day of the week.
Down the left side of the Calendar Heat Chart, the hour of the day is indicated. You can change the time format (either 12-hour or 24-hour) from the View Settings dialog.
Below each calendar chart is a list containing Filter options. The default filter, Show All, uses the available data, which is rated for each hour of the preceding 30 days, to fill the calendar grid with color. Tooltips provide the underlying values.
You can apply the following CA pattern-recognition algorithms to the chart to highlight patterns in the data:
Filters the Calendar Heat Chart view so that only the single busiest hour of each day is rated. The "busy hour" is the hour when utilization, latency, or loss reaches its highest average value for each hour. Use this filter to see the hour when demands on the device or interface were at their peak. The Busy Hour is reported for each day, whether the hour with the highest heat rating for that day was Green, Yellow, Orange, or Red.
Filters the view using the Red rating to find hourly patterns that are common across three or more days of each business week. If the same hour of three or more business-week days is rated Red (severely degraded or highly utilized), all of the days in the business week for that hour retain their ratings. All other hours not matching the pattern are changed to white. A business week runs Monday - Friday.
Filters the view using the Red rating to find hourly patterns that are common across four or more days of each calendar week. If the same hour of four or more calendar-week days is rated Red (severely degraded or highly utilized), all of the days in the calendar week for that hour retain their ratings. All other hours not matching the pattern are changed to white. A calendar week runs Sunday - Saturday.
Filters the view to find time-of-day patterns in business weeks, defined by default as Monday - Friday. If the same hour of the same business day is rated Red across 3 or more business weeks in the month, that hour of that same weekday retains its rating (Red, Orange, Yellow, or Green). The pattern applies across all matching business days in the month. All hours that do not match this pattern are changed to white.
Note: You can change the way a business week is defined by editing the view.
Filters the view to find time-of-day patterns in calendar weeks, defined as Sunday - Saturday. If the same hour of the same day is rated Red across 3 or more calendar weeks in the month, that hour of that same weekday retains its rating (Red, Orange, Yellow, or Green). The pattern applies across all matching days in the month. All hours that do not match this pattern are changed to white.
Similar logic applies to a calendar chart with a narrower timeframe, such as one week.
Note: A gray color indicates that no data was available for the indicated hour, and thus the hour was unrated.
The following table summarizes the default settings for calendar chart "heat" indicators:
|
Color |
Latency (ms) |
Utilization (%) |
Availability |
Loss (%) |
|---|---|---|---|---|
|
Green |
0 - 199 |
0 - 49 |
100 - 94 |
0 - 1 |
|
Yellow |
200 - 219 |
50 - 59 |
93 - 84 |
2 - 3 |
|
Orange |
220 - 239 |
60 - 69 |
83 - 1 |
4 - 5 |
|
Red |
240 and higher |
70 and higher |
0 |
6 and higher |
Note: Availability is an average of the following values: 0 (Unavailable); 100 (Available).
Most default values come from industry standards for user-impacting performance. However, the default values for the Availability metric are defined to emphasize down time and reveal how much down time occurred during the one-hour measurement period.
Acceptable values for round-trip latency depend on many variables, including transmission speed (circuit bandwidth), distance, and the latency sensitivity of particular applications. For example, a consistent round-trip latency of 300 ms across a satellite link may fall within acceptable bounds and meet SLAs. However, a consistent round-trip latency of 15 ms across a data center LAN could be a cause for concern about uplink capacity among Data Center switches. In this context, the smaller value would almost certainly fall outside of acceptable SLAs for the data center LAN. The same would be true for the Utilization and Loss metrics. Acceptable values vary by the supported technology.
|
Copyright © 2014 CA.
All rights reserved.
|
|