SysOrb Network Monitoring System Administrator's Guide: For version 3.6.0 | ||
---|---|---|
Prev | Chapter 12. SysOrb Tool | Next |
sysorb-tool has 3 different CSV formats, depending on which kind of check is selected.
The first line of each CSV file identifies the columns of the file. After this line, comes the real data from the check. Each line corresponds to one record in the SysOrb database. A record can contain several measurements. This is because SysOrb will "compress" older data, by combining several measurements into one record.
A measurement can be either good or bad. If it is good, it means that SysOrb actually have obtained a value from the check. If it is bad, it means that SysOrb was unable to obtain a value from the check. For example this is the case when SysOrb tries to perform a HTTP check, but cannot connect to the server.
Continuous checks are the checks that are plotted as a line graph in the SysOrb Web-interface, plus the uptime graphs. E.g. all NetChecks are Continuous graphs.
A CSV file of a continuous check contains 7 different columns.
The time at which the first measurement contained in the record is started.
The end time of the last measurement contained in the record.
The minimum value of the good measurements.
The average value of the good measurements included in the record.
The maximum value of the good measurements.
The number of good measurements contained in the record.
The number of bad measurements contained in the record.
State check are the checks that can be in different states. E.g. Process Presence check, RAID state check etc. are all State checks.
Internally in SysOrb the state of these checks are represented by a bit-mask. So when SysOrb "compresses" several state measurements into one, more bits get set in the bit-mask of the resulting record.
A state check CSV file contains 5 different fields:
The time at which the first measurement contained in the record is started.
The end time of the last measurement contained in the record.
A bitmask where all the bits of the different states seen in the period of the record are seen.
The number of good measurements contained in the record.
The number of bad measurements contained in the record.
String checks are all the kind of checks that produce strings. Currently Log Checks and Incident logs for Nodes are the only kinds.
A CSV file of a string check contains the 4 different fields:
The time at which the string was encountered.
An indicator of how severe the text message is. Currently three values are supported:
1 - Good
2 - Warning
4 - Alert
The text of the measurement.