| Mercury Monitor Reference Manual |
|---|
This section contains the metrics that are currently standardized.
The name of every host metric must start with the string host. Every metric in this category must have the following parameters:
| string host | The host name where the metric should be measured. A local host sensor must refuse to serve the metric if the value of this parameter does not resolve to an address belonging to a local network interface. If the specified host name resolves to more than one network addresses, at least one of them must be local. |
Host sensors should try to provide as many information as possible. If some information could not be determined exactly but a reasonable estimate can be made, this estimate should be published.
Most CPU-related metrics have the following optional parameter:
| uint32 cpu (optional) | The identifier of the CPU the consumer is interested in. The range of valid CPU identifiers is host specific (it may not start from 0 and it may not be continuous). |
Parameters:
| string host | The host name |
Data type: uint32 cpus
Measurement unit: pieces
Description: total number of CPUs installed on the system including ones that are not active at the moment. Note that if there are multiple logical CPUs implemented by the same piece of hardware, the returned value is the number of logical CPUs.
Parameters:
| string host | The host name |
Data type: uint32 cpus
Measurement unit: pieces
Description: number of CPUs that the operating system utilizes currently. Note that the operating system might reserve CPUs for special purposes so not all the reported CPUs might be available for running user processes.
Parameters:
| string host | The host name |
Data type: uint32 cpus
Measurement unit: pieces
Description: number of CPUs that are available for running general user processes.
Parameters:
| string host | The host name |
Data type: uint32 indexes[];
Description: array of valid CPU indexes. These indexes can be used as values to the "cpu" argument of various CPU metrics.
Parameters:
| string host | The host name |
Data type: uint32 freq
Measurement unit: Hz
Description: resolution of the time values returned by CPU usage metrics (host.cpu.user, host.cpu.system etc.). This value usually has nothing to do with the speed of the CPU.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: uint32 freq
Measurement unit: MHz
Description: operating frequency of a CPU. If the system supports mutliple processors running with different speeds, the cpu parameter selects which CPUs frequency will be reported. If the cpu parameter is omitted and there are CPUs running with different frequencies, the returned value will be the frequency of the first online CPU.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: string cpu_type
Description: a string describing the CPU type. The exact contents and format is architecture specific.
Parameters:
| string host | The host name |
| uint32 cpu | The CPU's id |
Data type: int32 cpu_state (enum: "Normal", "Offline", "Error", "PowerOff", "OS only", "User only", "Restricted");
Description: the current CPU state.
Parameters:
| string host | The host name |
| uint32 cpu | The CPU's id |
Data type: uint32 timestamp
Description: timestamp when the CPU entered its current state.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: uint32 ticks
Measurement unit: clockticks (see host.cpu.clocktick)
Description: time spent by the processor in user context. If the cpu parameter is omitted, the sum for all CPUs is returned.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: uint32 ticks
Measurement unit: clockticks (see host.cpu.clocktick)
Description: time spent by the processor servicing system calls. If the cpu parameter is omitted, the sum for all CPUs is returned.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: uint32 ticks
Measurement unit: clockticks (see host.cpu.clocktick)
Description: time spent by the processor waiting for external I/O to complete. If the cpu parameter is omitted, the sum for all CPUs is returned.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: uint32 ticks
Measurement unit: clockticks (see host.cpu.clocktick)
Description: time spent by the processor doing nothing. If the cpu parameter is omitted, the sum for all CPUs is returned.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: uint32 ticks
Measurement unit: clockticks (see host.cpu.clocktick)
Description: time spent by the processor servicing hardware interrupts. If the cpu parameter is omitted, the sum for all CPUs is returned.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: uint32 ticks
Measurement unit: clockticks (see host.cpu.clocktick)
Description: time spent by the processor running niced processes. If the cpu parameter is omitted, the sum for all CPUs is returned.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: uint32 ticks
Measurement unit: clockticks (see host.cpu.clocktick)
Description: time spent by the processor servicing soft interrupts. If the cpu parameter is omitted, the sum for all CPUs is returned.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: rec { uint32 user; uint32 system; uint32 iowait; uint32 idle }
Measurement unit: clockticks (see host.cpu.clocktick)
Description: Compound metrics for CPU usage. If the host supports the host.cpu.irq and/or host.cpu.softirq metrics, they are included in the system time. If the cpu parameter is omitted, the sum for all CPUs is returned.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: uint32 size
Measurement unit: Byte
Description: size of the 1st level data cache of the CPU.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: uint32 size
Measurement unit: Byte
Description: size of the 1st level instruction cache of the CPU.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: uint32 size
Measurement unit: Byte
Description: size of the 2nd level cache. If multiple processors share the same L2 cache the full cache size will be reported for each processor.
Parameters:
| string host | The host name |
| uint32 cpu (optional) | The CPU's id |
Data type: uint32 size
Measurement unit: Byte
Description: size of the 3rd level cache. If multiple processors share the same L3 cache the full cache size will be reported for each processor.
Parameters:
| string host | The host name |
Data type: rec { string dev; string mountpoint; } filesystems;
Description: list of the mounted filesystems.
Parameters:
| string host | The host name |
| string fs | Mount point |
Data type: rec { uint64 size (unit: "KiB"); uint64 avail (unit: "KiB"); uint64 reserved (unit: "KiB"); } space;
Description: size of the total/available/reserved space in a filesystem. Reserved means that only the superuser is allowed to use this space.
Parameters:
| string host | The host name |
| string fs | Mount point |
Data type: rec { uint64 total; uint64 avail; uint64 reserved; } inodes;
Description: number of total/available/reserved inodes in a filesystem. Reserved means that only files created by the superuser may use these inodes.
Parameters:
| string host | The host name |
Data type: uint64 (unit: KiB)
Description: size of the host's total physical memory usable by the operating system, in kilobytes. On some systems this may be smaller than the total amount of physical memory installed if there are reserved memory areas (like AGP aperture, hardware I/O buffers, firmware configuration information or the kernel's code itself).
Parameters:
| string host | The host name |
Data type: uint64 (unit: KiB)
Description: size of the host's free physical memory, in kilobytes. On many systems this metric is rather meaningless because the operationg system uses memory not allocated to running processes for caching and buffering purposes. On those systems (e.g. Linux), "free memory" means "wasted memory that you could remove from the system without performance degradation".
Parameters:
| string host | The host name |
Data type: uint64 (unit: KiB)
Description: size of memory that is shared between at least two processes on the system.
Parameters:
| string host | The host name |
Data type: uint64 (unit: KiB)
Description: size of memory that is currently being used for raw I/O buffers.
Parameters:
| string host | The host name |
Data type: uint64 (unit: KiB)
Description: size of memory that is currently being used for cached files.
Most network-interface metrics have the following parameter:
| string interface | The name of the network interface. |
Parameters:
| string host | The host name |
Data type: string name[];
Description: name of all the defined network interfaces (including ones that are not running currently).
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint64 amount (unit: Bytes)
Description: amount of data received on a network interface since it is defined. If the interface parameter is omitted, the sum over all up, running, non-loopback interfaces is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint64 count
Description: number of packets received on a network interface since it is defined. If the interface parameter is omitted, the sum over all up, running, non-loopback interfaces is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint32 count
Description: number of bad packets received on a network interface since it is defined. If the interface parameter is omitted, the sum over all up, running, non-loopback interfaces is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint32 count
Description: number of received packets dropped (usually due to resource shortage). If the interface parameter is omitted, the sum over all up, running, non-loopback interfaces is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint64 amount (unit: Bytes)
Description: amount of data sent out on a network interface since it is defined. If the interface parameter is omitted, the sum over all up, running, non-loopback interfaces is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint64 count
Description: number of packets sent out on a network interface since it is defined. If the interface parameter is omitted, the sum over all up, running, non-loopback interfaces is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint32 count
Description: number of transmission problems on a network interface since it is defined. If the interface parameter is omitted, the sum over all up, running, non-loopback interfaces is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint32 count
Description: number of to-be-sent packets dropped (usually due to resource shortage). If the interface parameter is omitted, the sum over all up, running, non-loopback interfaces is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint32 count
Description: number of carrier failures on a network interface. If the interface parameter is omitted, the sum over all up, running, non-loopback interfaces is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint32 count
Description: number of collisions detected by the interface. If the interface parameter is omitted, the sum over all up, running, non-loopback interfaces is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint64 amount (unit: Bytes)
Description: amount of data sent or received on a network interface since it is defined. If the interface parameter is omitted, the sum over all up, running, non-loopback interfaces is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint64 count
Description: number of packets sent or received on a network interface since it is defined. If the interface parameter is omitted, the sum over all up, running, non-loopback interfaces is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint32 count
Description: number of bad packets sent or received on a network interface since it is defined. If the interface parameter is omitted, the sum over all up, running, non-loopback interfaces is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: uint32 mtu
Description: MTU size of the interface. If the interface parameter is omitted, the value for the first up, running, non-loopback interface is returned.
Parameters:
| string host | The host name |
| string interface | The network interface name |
Data type: uint32 flags (flags: "Up", "Broadcast", "PoinToPoint", "Loopback", "Running", "Promisc", "Multicast")
Description: interface flags.
Parameters:
| string host | The host name |
| string interface | The network interface name |
Data type: uint32 index
Description: the internal index of an interface. Some systems allow renaming network interfaces on the fly, but the interface index remains constant while the interface is defined.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: rec { uint32 addr (format: ipv4); uint32 netmask (format: ipv4); uint32 broadcast (format: ipv4); uint32 dest (format: ipv4); string label; } addresses[];
Description: the IPv4 address(es) the interface has. If the address is attached to the interface directly, the label field will be empty. If the address is attached as an interface alias, the label field will contain the name of the interface alias. If the interface parameter is omitted, the value for the first up, running, non-loopback interface is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: rec { opaque addr (format: ipv6); uint32 prefixlen; opaque dest (format: ipv6); string label; } addresses[];
Description: the IPv6 address(es) the interface has. If the address is attached to the interface directly, the label field will be empty. If the address is attached as an interface alias, the label field will contain the name of the interface alias. If the interface parameter is omitted, the value for the first up, running, non-loopback interface is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: opaque hwaddr (format: hwaddr);
Description: the hardware address of the interface. If the interface parameter is omitted, the value for the first up, running, non-loopback interface is returned.
Parameters:
| string host | The host name |
| string interface (optional) | The network interface name |
Data type: string hwtype;
Description: the hardware type of the interface. If the interface parameter is omitted, the value for the first up, running, non-loopback interface is returned.
Parameters:
| string host | The host name |
Data type: string osname;
Description: name of the operating system.
Parameters:
| string host | The host name |
Data type: string release;
Description: the release identifier of the operating system. The exact contents and meaning is operating system dependent.
Parameters:
| string host | The host name |
Data type: string version;
Description: version of the operating system. The format of the version string is operating system dependent.
Parameters:
| string host | The host name |
Data type: string machine;
Description: system architecture identifier.
Parameters:
| string host | The host name |
Data type: uint32 timestamp (format: unixtime);
Description: time when the currently running operating system instance booted.
Parameters:
| string host | The host name |
Data type: uint64 (unit: KiB)
Description: total size of the available swap area(s). This includes any active swap files and/or swap partitions.
Parameters:
| string host | The host name |
Data type: uint64 (unit: KiB)
Description: size of the swap area that is currently unused.
Parameters:
| string host | The host name |
Data type: uint64 (unit: KiB)
Description: amount of memory that was read in from the swap area(s) since boot.
Parameters:
| string host | The host name |
Data type: uint64 (unit: KiB)
Description: amount of memory that was written to the swap area(s) since boot.
Parameters:
| string host | The host name |
Data type: uint64 (unit: KiB)
Description: amount of memory that was read since boot.
Parameters:
| string host | The host name |
Data type: uint64 (unit: KiB)
Description: amount of memory that was written since boot.
Parameters:
| string host | The host name |
Data type: uint64 count
Description: number of page faults since boot.
Parameters:
| string host | The host name |
Data type: uint64 count
Description: number of major page faults since boot.
Parameters:
| string host | The host name |
Data type: rec { string device; uint64 size (unit: "KiB"); } disks[];
Description: name and size of disk devices connected to the system
Parameters:
| string host | The host name |
| string disk (optional) | The host name |
Data type: rec { string device; uint64 size (unit: "KiB"); } disks[];
Description: name and size of disk partitions. If disk is not given, partitions on all known disks are returned.
Parameters:
| string host | The host name |
| string disk (optional) | The host name |
Data type: rec { uint32 reads; uint32 writes; uint64 read_kb (unit: "KiB"); uint64 write_kb (unit: "KiB"); uint32 in_progress; uint64 time (unit: "ms"); uint64 weighted_time (unit: "ms") } stats;
Description: I/O statistics. If disk is missing, the totals over all disks are returned.
Parameters:
| string jobid | The job ID |
| string progname (optional) | The process name |
| int32 tid (optional) | The thread ID |
Data type: string
Description: generic application-generated message.
Parameters:
| string jobid | The job ID |
| string progname (optional) | The process name |
| int32 tid (optional) | The thread ID |
Data type: rec { string progname; int32 tid; string event; } event;
Description: information about processes starting, quitting or crashing.
Parameters:
| string jobid | The job ID |
| string progname (optional) | The process name |
| int32 tid (optional) | The thread ID |
| string name | The name of the variable |
| int32 start (optional) | The index of the first array element to return (default: 0) |
| int32 length (optional) | The number of array elements to return (default: all) |
Data type: rec { string progname; int32 tid; int32 value[]; } event;
Description: get the value of an application-defined array of 32-bit integers.
Parameters:
| string jobid | The job ID |
| string progname (optional) | The process name |
| int32 tid (optional) | The thread ID |
| string name | The name of the variable |
| int32 start (optional) | The index of the first array element to return (default: 0) |
| int32 length (optional) | The number of array elements to return (default: all) |
Data type: rec { string progname; int32 tid; int64 value[]; } value;
Description: get the value of an application-defined array of 64-bit integers.
Parameters:
| string jobid | The job ID |
| string progname (optional) | The process name |
| int32 tid (optional) | The thread ID |
| string name | The name of the variable |
| int32 start (optional) | The index of the first array element to return (default: 0) |
| int32 length (optional) | The number of array elements to return (default: all) |
Data type: rec { string progname; int32 tid; double value[]; } value;
Description: get the value of an application-defined array of 64-bit double precision floating point numbers.
Parameters:
| string jobid | The job ID |
| string progname (optional) | The process name |
| int32 tid (optional) | The thread ID |
| string name | The name of the variable |
| int32 start (optional) | The index of the first character of the string to return (default: 0) |
| int32 length (optional) | The number of characters to return (default: all) |
Data type: rec { string progname; int32 tid; string value; } value;
Description: get the value of an application-defined string (or part of it).
Parameters:
| string jobid | The job ID |
| string progname (optional) | The process name |
| int32 tid (optional) | The thread ID |
| string name | The name of the variable |
| int32 start (optional) | The index of the first byte of the variable to return (default: 0) |
| int32 length (optional) | The number of bytes to return (default: all) |
Data type: rec { string progname; int32 tid; opaque value; } value;
Description: get the value of an application-defined opaque variable (or part of it).
Parameters:
| string jobid | The job ID |
| string progname (optional) | The process name |
| int32 tid (optional) | The thread ID |
| string name | The name of the variable |
Data type: rec { string progname; int32 tid; string type; int32 length; boolean readonly; } info;
Description: get information about an application-defined variable
Parameters:
| string name | The name of the control |
Data type: rec { int32 domain; string name; string type; rec { string name; string type; uint32 flags; } params[]; int32 measurement; } definition;
Description: definition of the specified control.
Parameters:
| string name | The name of the metric |
Data type: rec { int32 domain; string name; string type; rec { string name; string type; uint32 flags; } params[]; int32 measurement; } definition;
Description: definition of the specified metric.
Data type: uint32 metric;
Description: this special metric is sent when a metric instance is destroyed inside the producer.
Data type: rec { uint32 metric; uint32 count (unit: "messages"); } value;
Description: this special metric is sent when the producer had to drop some messages due to the length of the message queue exceeding the allowed limit.
| << Monitoring Protocol | GridLab Project >> |