| 1 |
+ |
/* |
| 2 |
+ |
* i-scream central monitoring system |
| 3 |
+ |
* http://www.i-scream.org.uk |
| 4 |
+ |
* Copyright (C) 2000-2002 i-scream |
| 5 |
+ |
* |
| 6 |
+ |
* This program is free software; you can redistribute it and/or |
| 7 |
+ |
* modify it under the terms of the GNU General Public License |
| 8 |
+ |
* as published by the Free Software Foundation; either version 2 |
| 9 |
+ |
* of the License, or (at your option) any later version. |
| 10 |
+ |
* |
| 11 |
+ |
* This program is distributed in the hope that it will be useful, |
| 12 |
+ |
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 13 |
+ |
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 14 |
+ |
* GNU General Public License for more details. |
| 15 |
+ |
* |
| 16 |
+ |
* You should have received a copy of the GNU General Public License |
| 17 |
+ |
* along with this program; if not, write to the Free Software |
| 18 |
+ |
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. |
| 19 |
+ |
*/ |
| 20 |
+ |
|
| 21 |
|
//---PACKAGE DECLARATION--- |
| 22 |
< |
package uk.ac.ukc.iscream.client.monitors; |
| 22 |
> |
package uk.org.iscream.cms.server.client.monitors; |
| 23 |
|
|
| 24 |
|
//---IMPORTS--- |
| 25 |
|
import java.util.HashMap; |
| 26 |
< |
import uk.ac.ukc.iscream.client.*; |
| 27 |
< |
import uk.ac.ukc.iscream.core.*; |
| 28 |
< |
import uk.ac.ukc.iscream.util.*; |
| 29 |
< |
import uk.ac.ukc.iscream.componentmanager.*; |
| 26 |
> |
import uk.org.iscream.cms.server.client.*; |
| 27 |
> |
import uk.org.iscream.cms.server.core.*; |
| 28 |
> |
import uk.org.iscream.cms.server.util.*; |
| 29 |
> |
import uk.org.iscream.cms.server.componentmanager.*; |
| 30 |
|
|
| 31 |
|
/** |
| 32 |
|
* This Monitor watches the CPU load for all machines |
| 34 |
|
* @author $Author$ |
| 35 |
|
* @version $Id$ |
| 36 |
|
*/ |
| 37 |
< |
public class CPU__Monitor implements PluginMonitor { |
| 37 |
> |
public class CPU__Monitor extends MonitorSkeleton { |
| 38 |
|
|
| 39 |
|
//---FINAL ATTRIBUTES--- |
| 40 |
|
|
| 43 |
|
*/ |
| 44 |
|
public final String REVISION = "$Revision$"; |
| 45 |
|
|
| 46 |
+ |
/** |
| 47 |
+ |
* A description of this monitor |
| 48 |
+ |
*/ |
| 49 |
|
public final String DESC = "Monitors CPU."; |
| 50 |
|
|
| 51 |
|
//---STATIC METHODS--- |
| 53 |
|
//---CONSTRUCTORS--- |
| 54 |
|
|
| 55 |
|
//---PUBLIC METHODS--- |
| 56 |
< |
|
| 56 |
> |
|
| 57 |
> |
/** |
| 58 |
> |
* Analyse a packet of data, and generate an alert if |
| 59 |
> |
* necessary. |
| 60 |
> |
* |
| 61 |
> |
* @param packet the XMLPacket to analyse |
| 62 |
> |
*/ |
| 63 |
|
public void analysePacket(XMLPacket packet) { |
| 64 |
< |
if (packet.getParam("packet.attributes.type").equals("data")) { |
| 65 |
< |
String source = packet.getParam("packet.attributes.machine_name"); |
| 66 |
< |
if (!_hosts.containsKey(source)) { |
| 67 |
< |
_hosts.put(source, new Register(source, _name, _attributes.length)); |
| 68 |
< |
} |
| 40 |
< |
|
| 41 |
< |
Register reg = (Register) _hosts.get(source); |
| 42 |
< |
for(int attributeNum = 0; attributeNum < _attributes.length; attributeNum++) { |
| 43 |
< |
// find out the threshold level we're at |
| 44 |
< |
int result = checkAttributeThreshold(packet.getParam(_attributes[attributeNum]), reg); |
| 45 |
< |
|
| 46 |
< |
// decide what threshold level we're on, if we've changed, record that |
| 47 |
< |
if (result != reg.getLastThresholdLevel(attributeNum)) { |
| 48 |
< |
reg.setLastThresholdLevel(attributeNum, result); |
| 49 |
< |
} |
| 50 |
< |
// as long as this isn't a normal level |
| 51 |
< |
if(reg.getLastThresholdLevel(attributeNum) != Alert.thresholdNORMAL) { |
| 52 |
< |
// if the time since the last alert is more than the time for |
| 53 |
< |
// its timeout, fire an alert, escalate the alert |
| 54 |
< |
long timeout = reg.getAlertTimeout(reg.getLastAlertLevel(attributeNum), attributeNum); |
| 55 |
< |
if ((timeout > 0) && (reg.getTimeLastSent(attributeNum) > 0)) { |
| 56 |
< |
if((System.currentTimeMillis() - reg.getTimeLastSent(attributeNum)) > timeout) { |
| 57 |
< |
reg.escalateAlert(attributeNum); |
| 58 |
< |
reg.setTimeLastSent(attributeNum, System.currentTimeMillis()); |
| 59 |
< |
fireAlert(reg, packet, attributeNum); |
| 60 |
< |
} |
| 61 |
< |
// if we don't have a timeout configured...we got STRAIGHT to the next level |
| 62 |
< |
} else { |
| 63 |
< |
reg.escalateAlert(attributeNum); |
| 64 |
< |
reg.setTimeLastSent(attributeNum, System.currentTimeMillis()); |
| 65 |
< |
fireAlert(reg, packet, attributeNum); |
| 66 |
< |
} |
| 67 |
< |
|
| 68 |
< |
// we must be on ok, check the timeout value for this |
| 69 |
< |
} else { |
| 70 |
< |
// if we were on an OK alert before, then we don't do anything |
| 71 |
< |
// but if we weren't we only set OK, once the timout of the last |
| 72 |
< |
// alert has occourd |
| 73 |
< |
if (reg.getLastAlertLevel(attributeNum) != Alert.alertOK) { |
| 74 |
< |
long timeout = reg.getAlertTimeout(reg.getLastAlertLevel(attributeNum), attributeNum); |
| 75 |
< |
if ((timeout > 0) && (reg.getTimeLastSent(attributeNum) > 0)) { |
| 76 |
< |
if ((System.currentTimeMillis() - reg.getTimeLastSent(attributeNum)) > timeout) { |
| 77 |
< |
reg.setLastAlertLevel(attributeNum, Alert.alertOK); |
| 78 |
< |
reg.setTimeLastSent(attributeNum, System.currentTimeMillis()); |
| 79 |
< |
fireAlert(reg, packet, attributeNum); |
| 80 |
< |
} |
| 81 |
< |
} else { |
| 82 |
< |
reg.setLastAlertLevel(attributeNum, Alert.alertOK); |
| 83 |
< |
reg.setTimeLastSent(attributeNum, System.currentTimeMillis()); |
| 84 |
< |
fireAlert(reg, packet, attributeNum); |
| 85 |
< |
} |
| 86 |
< |
} |
| 87 |
< |
} |
| 88 |
< |
} |
| 64 |
> |
String source = packet.getParam("packet.attributes.machine_name"); |
| 65 |
> |
if (!_hosts.containsKey(source)) { |
| 66 |
> |
HashMap attributeRegisters = new HashMap(); |
| 67 |
> |
initAttributeRegsiters(source, attributeRegisters); |
| 68 |
> |
_hosts.put(source, attributeRegisters); |
| 69 |
|
} |
| 70 |
+ |
|
| 71 |
+ |
HashMap attributeRegisters = (HashMap) _hosts.get(source); |
| 72 |
+ |
for(int attributeNum = 0; attributeNum < _attributes.length; attributeNum++) { |
| 73 |
+ |
Register reg = (Register) attributeRegisters.get(_attributes[attributeNum]); |
| 74 |
+ |
// find out the threshold level we're at |
| 75 |
+ |
String attribute = _attributes[attributeNum]; |
| 76 |
+ |
String attributeName = _attributeNames[attributeNum]; |
| 77 |
+ |
String currentValue = packet.getParam(attribute); |
| 78 |
+ |
int newThreshold = checkAttributeThreshold(currentValue, reg); |
| 79 |
+ |
processAlert(newThreshold, attributeName, reg, source, currentValue); |
| 80 |
+ |
} |
| 81 |
|
} |
| 82 |
|
|
| 83 |
|
/** |
| 84 |
|
* Overrides the {@link java.lang.Object#toString() Object.toString()} |
| 85 |
|
* method to provide clean logging (every class should have this). |
| 86 |
|
* |
| 87 |
< |
* This uses the uk.ac.ukc.iscream.util.NameFormat class |
| 87 |
> |
* This uses the uk.org.iscream.cms.server.util.NameFormat class |
| 88 |
|
* to format the toString() |
| 89 |
|
* |
| 90 |
|
* @return the name of this class and its CVS revision |
| 104 |
|
} |
| 105 |
|
|
| 106 |
|
//---PRIVATE METHODS--- |
| 107 |
< |
|
| 107 |
> |
|
| 108 |
> |
/** |
| 109 |
> |
* Checks a piece of current data, and returns the |
| 110 |
> |
* threshold it breaches, if any. |
| 111 |
> |
* |
| 112 |
> |
* @param attributeString a String representing the current data value |
| 113 |
> |
* @param reg the Register for the host |
| 114 |
> |
* @return the threshold level breached, if any |
| 115 |
> |
*/ |
| 116 |
|
private int checkAttributeThreshold(String attributeString, Register reg) { |
| 117 |
|
for(int thresholdLevel = Alert.thresholdLevels.length - 1; thresholdLevel >= 0; thresholdLevel--) { |
| 118 |
|
if (reg.getThreshold(thresholdLevel) != -1.0) { |
| 127 |
|
} |
| 128 |
|
} |
| 129 |
|
} |
| 130 |
< |
return 0; |
| 130 |
> |
return Alert.thresholdNORMAL; |
| 131 |
|
} |
| 132 |
< |
|
| 133 |
< |
private void fireAlert(Register reg, XMLPacket packet, int attributeNum) { |
| 134 |
< |
int alertLevel = reg.getLastAlertLevel(attributeNum); |
| 135 |
< |
int thresholdLevel = reg.getLastThresholdLevel(attributeNum); |
| 136 |
< |
String source = packet.getParam("packet.attributes.machine_name"); |
| 137 |
< |
String currentValue = packet.getParam(_attributes[attributeNum]); |
| 138 |
< |
String attributeName = _attributeNames[attributeNum]; |
| 139 |
< |
String thresholdValue = Double.toString(reg.getThreshold(thresholdLevel)); |
| 140 |
< |
String time = Long.toString(reg.getAlertTimeout(reg.getLastAlertLevel(attributeNum), attributeNum) / 1000); |
| 141 |
< |
if (thresholdLevel == Alert.thresholdNORMAL) { |
| 142 |
< |
thresholdValue = "-"; |
| 132 |
> |
|
| 133 |
> |
/** |
| 134 |
> |
* Initialises a HashMap of Registers with the current list |
| 135 |
> |
* of attributes. This is only used if we are looking at more |
| 136 |
> |
* than one distinct attribute. |
| 137 |
> |
* |
| 138 |
> |
* @param source the host we are looking at |
| 139 |
> |
* @param attributeRegisters a HashMap to put the new Registers in |
| 140 |
> |
*/ |
| 141 |
> |
private void initAttributeRegsiters(String source, HashMap attributeRegisters) { |
| 142 |
> |
for(int attributeNum = 0; attributeNum < _attributes.length; attributeNum++) { |
| 143 |
> |
String attributeName = _attributes[attributeNum].substring(_attributes[attributeNum].lastIndexOf(".") + 1); |
| 144 |
> |
attributeRegisters.put(_attributes[attributeNum], new Register(source, _name, attributeName)); |
| 145 |
|
} |
| 145 |
– |
if (alertLevel == Alert.alertOK) { |
| 146 |
– |
time = "0"; |
| 147 |
– |
} |
| 148 |
– |
Alert alert = new Alert(alertLevel, thresholdLevel, source, thresholdValue, currentValue, attributeName, time); |
| 149 |
– |
_alerterQueue.add(alert); |
| 150 |
– |
_logger.write(toString(), Logger.DEBUG, "Fired alert for source:" + source + " at alert level:" + Alert.alertLevels[alertLevel] + " on:" + attributeName + " for threshold level:" + Alert.thresholdLevels[thresholdLevel] + " at:" + currentValue + " exceeding threshold of:" +thresholdValue + " next alert sent in:" + time + "secs"); |
| 146 |
|
} |
| 147 |
|
|
| 148 |
|
//---ACCESSOR/MUTATOR METHODS--- |
| 149 |
+ |
|
| 150 |
+ |
/** |
| 151 |
+ |
* Returns a reference to a specific Queue for this |
| 152 |
+ |
* monitor. This Queue returns only the data packets |
| 153 |
+ |
* (based on type) that we want too look at. |
| 154 |
+ |
* |
| 155 |
+ |
* @return a reference to a Queue |
| 156 |
+ |
*/ |
| 157 |
+ |
protected Queue getQueue() { |
| 158 |
+ |
return MonitorManager.getInstance().getDataQueue(); |
| 159 |
+ |
} |
| 160 |
|
|
| 161 |
|
//---ATTRIBUTES--- |
| 162 |
|
|
| 170 |
|
* be changed to null for utility classes. |
| 171 |
|
*/ |
| 172 |
|
private String _name = "CPU"; |
| 173 |
< |
|
| 173 |
> |
|
| 174 |
|
/** |
| 175 |
< |
* This holds a reference to the |
| 176 |
< |
* system logger that is being used. |
| 175 |
> |
* A HashMap of Registers (or groups of Registers), one |
| 176 |
> |
* for each host we're monitoring. |
| 177 |
|
*/ |
| 178 |
< |
private Logger _logger = ReferenceManager.getInstance().getLogger(); |
| 178 |
> |
private HashMap _hosts = new HashMap(); |
| 179 |
|
|
| 174 |
– |
private Queue _alerterQueue = ClientMain._alerterQueue; |
| 175 |
– |
|
| 180 |
|
/** |
| 181 |
< |
* A reference to the configuration proxy in use |
| 181 |
> |
* An array of attributes which we will be checking. |
| 182 |
|
*/ |
| 179 |
– |
private ConfigurationProxy _cp = ConfigurationProxy.getInstance(); |
| 180 |
– |
|
| 181 |
– |
private HashMap _hosts = new HashMap(); |
| 182 |
– |
|
| 183 |
|
private String[] _attributes = { "packet.cpu.user", "packet.cpu.kernel", "packet.cpu.iowait", "packet.cpu.swap" }; |
| 184 |
+ |
|
| 185 |
+ |
/** |
| 186 |
+ |
* An array of "nice names" for the attributes in _attributes. |
| 187 |
+ |
*/ |
| 188 |
|
private String[] _attributeNames = {"User CPU", "Kernel CPU", "I/O Wait CPU", "Swap CPU"}; |
| 189 |
|
|
| 190 |
|
//---STATIC ATTRIBUTES--- |
