toni-moreno / cellhealth-ng Goto Github PK

View Code? Open in Web Editor NEW

This project forked from buck2k/cellhealth-ng

5.0 5.0 1.0 74.51 MB

An OpenSource tool to get performance Metrics from WebSphere PMI and send them to a Graphite Server.

License: GNU General Public License v2.0

Shell 2.20% Java 97.80%

graphite influxdb metrics monitoring performance pmi websphere

cellhealth-ng's Introduction

Hi there 👋

Thanks for visiting my GitHub profile, it's great to meet you here! 😊

Hi, I'm Toni Moreno from Barcelona.

🌱 I’m currently learning everything about technology
🤔 I will learn programming Rust some day...
📫 The best way to reach me is by sending me an email.
💬 Ask me about anything here.

GitHub Stats:

cellhealth-ng's People

Contributors

Stargazers

Watchers

Forkers

rsolorzano2410

cellhealth-ng's Issues

Add JMX too?

Hi,

Any chance you adding JMX stats as well? Or at least point me in the right direct on how to add it based on your design?

Dots in instance names are not tranlated to "_"

When sendind data to the graphite backend new graphite levels appears depending on the dot number

CellHealth seems to be sending 1 write request per metric

Hi, trying InfluxDB configuration on a WAS 8.5.5 seems to be doing 400+ HTTP Request on a single period.

Number of writes on a single period:

#grep was_metrics influxdb_httpaccess.log | grep POST | grep  12:09 | wc -l
424

Request example. The body size seems to be very small:

X.X.X.X.X - was_writer [10/Sep/2019:12:40:16 +0200] "POST /write?consistency=one&db=was_metrics&p=%5BREDACTED%5D&precision=n&rp=&u=was_writer HTTP/1.1" 204 0 "-" "influxdb-smart-relay" 60d67f0d-d3b7-11e9-b424-0050569975eb 5685
X.X.X.X.X - was_writer [10/Sep/2019:12:40:16 +0200] "POST /write?consistency=one&db=was_metrics&p=%5BREDACTED%5D&precision=n&rp=&u=was_writer HTTP/1.1" 204 0 "-" "influxdb-smart-relay" 60d78de1-d3b7-11e9-b425-0050569975eb 5160

Is it sending a write HTTP request by metric gathered?

Thanks,
Regards

Error in ch_stats.os.totalmemory

cellhealth in sending null value and graphite discard it as and invalid line.

Exception on getSelfStats() stops thread manager

Exception in thread "manager" java.lang.NullPointerException
        at java.util.LinkedList$ListItr.next(LinkedList.java:903)
        at cellhealth.core.threads.Metrics.ThreadManager.launchThreads(ThreadManager.java:112)
        at cellhealth.core.threads.Metrics.ThreadManager.run(ThreadManager.java:51)
        at java.lang.Thread.run(Thread.java:798)

CH can not reconnect if DMGR is restarted

After a WebSphere restart the cellhealth process can not reconnect itself.

Cellhealth-ng reconnects but doesn't send metrics

It seems like it can perform TCP/IP re-connection, first time the carbon server stops. The next server stop, can do TCP/IP but the metric gather loop seems freezed.

Error on SOAP timeout not handled

with buggy or slow SOAP connections the agent throws this exception and stops collecting metrics.

2017-03-13 19:15:23,262 [ERROR][5767578] manager |  [CellHealth 1.0] MBeansManager, Error in connector
com.ibm.websphere.management.exception.ConnectorException: ADMC0009E: The system failed to make the SOAP RPC call: queryNames
        at com.ibm.ws.management.connector.soap.SOAPConnectorClient.invokeTemplateOnce(SOAPConnectorClient.java:886)
        at com.ibm.ws.management.connector.soap.SOAPConnectorClient.invokeTemplate(SOAPConnectorClient.java:682)
        at com.ibm.ws.management.connector.soap.SOAPConnectorClient.invokeTemplate(SOAPConnectorClient.java:672)
        at com.ibm.ws.management.connector.soap.SOAPConnectorClient.queryNames(SOAPConnectorClient.java:584)
        at com.ibm.ws.management.connector.soap.SOAPConnectorClient.invoke(SOAPConnectorClient.java:509)
        at com.sun.proxy.$Proxy0.queryNames(Unknown Source)
        at com.ibm.ws.management.AdminClientImpl.queryNames(AdminClientImpl.java:108)
        at cellhealth.core.connection.MBeansManager.getMBeans(MBeansManager.java:49)
        at cellhealth.core.connection.MBeansManager.getAllServerRuntimes(MBeansManager.java:101)
        at cellhealth.core.threads.Metrics.ThreadManager.launchThreads(ThreadManager.java:71)
        at cellhealth.core.threads.Metrics.ThreadManager.run(ThreadManager.java:50)
        at java.lang.Thread.run(Thread.java:790)
Caused by: [SOAPException: faultCode=SOAP-ENV:Client; msg=Read timed out; targetException=java.net.SocketTimeoutException: Read timed out]
        at org.apache.soap.transport.http.SOAPHTTPConnection.send(SOAPHTTPConnection.java:479)
        at org.apache.soap.rpc.Call.WASinvoke(Call.java:451)
        at com.ibm.ws.management.connector.soap.SOAPConnectorClient$8.run(SOAPConnectorClient.java:844)
        at com.ibm.ws.security.util.AccessController.doPrivileged(AccessController.java:118)
        at com.ibm.ws.management.connector.soap.SOAPConnectorClient.invokeTemplateOnce(SOAPConnectorClient.java:837)
        ... 11 more
```

CH reports "null" in some metrics

This is the output

test.null.was.null.jvm.memoryPool.miscellaneous_non-heap_storage.init 0 1479475837
test.null.was.null.jvm.memoryPool.miscellaneous_non-heap_storage.used 0 1479475837
test.null.was.null.jvm.memoryPool.miscellaneous_non-heap_storage.committed 1353648 1479475837
test.null.was.null.jvm.memoryPool.miscellaneous_non-heap_storage.max -1 1479475837
test.null.was.null.jvm.memoryPool.Java_heap.init 4194304 1479475837
test.null.was.null.jvm.memoryPool.Java_heap.used 6275832 1479475837
test.null.was.null.jvm.memoryPool.Java_heap.committed 7929856 1479475837
test.null.was.null.jvm.memoryPool.Java_heap.max 536870912 1479475837
test.null.was.null.jvm.threads.daemon 9 1479475837
test.null.was.null.jvm.threads.thread_count 17 1479475837

seems to be metrics from itself

[Feature Request] Allow to set precision on InfluxDB write request

Hi, trying InfluxDB configuration on a WAS 8.5.5 and seems that is writing with a default precision of n.

I'm interested to store it with a s precision but it seems not to be able to set it via config.

Could be interesting to add field to set it up

[Feature Request] add configurable hostnames from nodenames.

for custom nodenames the current hostname parsing is not working , so we will need to set a new property from cellhealth in the config.properties to disable automatic hostname to nodename mapping and another in the graphite.properties so set a prefix mapping with the hostname from each nodename.