What tool do you use to monitor your servers?
For an extra thorough checklist of monitoring devices and also their attributes, look into this Wikipedia page.
As the inquiry states, what are one of the most generally made use of devices made use of for this job and also what are their toughness and also weak points?
ServersAlive is a reasonably economical, straightforward device for all type of ballot, consisting of TCP solutions, Windows solutions, your very own personalized manuscripts, whatever. The feedback from the programmer on his newsletter is quick and also individual.
I utilized it at a previous work for solution monitoring and also it was trusted, customisable and also economical.
MSP Center (the previous OpManager) is actually irritating to make use of and also I can not advise it. The user interface is totally internet - based which suggests absolutely no responses and also a randomly minimal set of selections at any time you intend to do something. Their internet site appears packed with pointers and also documents, yet it is a little bit like Outlook - it assures an entire number of power yet is hindered by some programmer is minimal creative imagination.
If you are seeking an absolutely no - config remedy for your helpdesk, well possibly, yet it is none type of power device. If you have time to tune your monitoring to fulfill your demands after that there are various other remedies that would certainly compensate your initiatives extra.
A new participant on the scene to look into for taking on Cacti and also the RRDTool based remedies is Graphite (http://graphite.wikidot.com/)
RRDTool is changed with a support store called Whisper. The docs offer a respectable review of why it varies and also I actually like the CLI for impromptu graphing when exploring something.
We required something customisable as we require to check some systems which are not on-line regularly, yet can send mail or be called in.
We attempted nagios (puzzle of manuscripts), AppManager (wonderful, yet nonadaptable), Zenoss (wonderful, yet when you state Oracle, rate obtains substantial multipliers) and also came down on Zabbix which has open method, open data source framework, hell, I can write a plugin on every degree in a hr.
It is perfectly compartmentalised (web server, customer, data source,
YMMV, for us the monitoring of "offline" systems is necessary and also it is generally not covered by such software program.
Nagios and also HPOpenview are both that I know with and also have experience in. Both are excellent selections, although for the last I'll resemble various other posters that it requires a person that recognizes just how to do it right. chicken once more the only area I saw it running was when I was with HP to make sure that could have aided my assumption.
We've attempted Applications Manager Its working on java and also mysql. It is actually effective and also very easy to set up from the internet browser. It is not that pricey either.
Presently we make use of SCOM from MS. I would not advise it to any person!
We make use of Ipswitch whatsup gold 12 for monitoring concerning 2000 tools, both efficiency and also tcp/ip or wmi based displays and also both windows and also linux. Good idea concerning it is that it is very easy to make use of and also set up, has bulk adjustment alternatives and also autodiscovery, numerous alert approaches. The negative side: appears to have had a restriction of concerning 2000 tools, afterwards efficiency was obtaining slow-moving, plus it just works on windows. The dispersed variation does not actually be entitled to the name and also the cost. We reviewed nagios (arrangement also intricate for a vibrant setting), zenoss (no mass adjustment or autodiscovery, also minimal for vibrant setting) and also presently considering Zabbix, which appears most encouraging with all the wonderful attributes Whatsup has and also extra, such as totally dispersed style with probes and also web server, reasonably straightforward arrangement, open source backend (mysql, apache)
I've been making use of Sysmon for a variety of years. There are a couple of modern-day solutions that it does not check, yet it assembles conveniently on the majority of * nix systems, has virtually no dependancies, is exceptionally light - weight, can check large varieties of tools and also solutions effortlessly, can take care of intricate network formats (incl. ring geographies) and also failover monitoring. It is primarily a config documents bargain, yet the layout is rather very easy (based upon plist/css).
We make use of AlertFox given that a couple of weeks and also are really satisfied it. It not just examines our uptime and also performance, yet additionally checks purchasing cart, customer login and also various other essential components of the internet site using purchase manuscripts (iMacros based).
For our inner monitoring (disk room etc) we make use of Nagios .
If you remain in a rush and also desire a fast device to check your MS web server after that make use of efficiency display for windows, set up a counter log with personalized monitoring layout and also a custome timetable (eg: accumulate information for 5 minutes every hr). After that download and install Microsoft is LogParser and also Codeplex is Performance Analysis of Logs (PAL) Tool (http://pal.codeplex.com/) to crisis your counter log. Chum will certainly create a wonderful recorded record with web links to feasible concern addressing documents/tools.
I've collaborated with a great deal of monitoring systems at a great deal of areas. A lot of them have actually currently been stated. Below are a couple of that have not been:
SMARTS - currently possessed by EMC. Actually is the most effective point ever before for origin. It is not economical and also assistance might not be excellent any longer as it is possessed by EMC. We were fortunate adequate to collaborate with the owners of the firm to get it applied.
Large Brother. Wonderful and also straightforward, yet a negative certificate. It is additionally the ugliest internet gui I've ever before seen, so I needed to revise it. Never ever obtained Big Sister to function.
HP Openview, when crafted, mounted and also run by an experienced designer can be excellent. Nonetheless I've just seen it done right as soon as and also wrong regularly than I can bear in mind. I would certainly never ever pick to utilize it.
BMC Patrol. Simply horrible. Pass away, pass away!
And also ultimately, for logs and also locating troubles you simply need to make use of Splunk. If this had actually been around 10 years ago I would certainly have conserved myself a great deal of lost time.
We make use of OpsView, which works on top of Nagios. The webUI aids us release new host display interpretations without needing to permit SSH accessibility, gives public sights, and also documents historic values. This comes in handy for provisioning and also establishing ideal standards.
I make use of Polymon and also enjoy it.
It is superb for monitoring anything that can be connected by TCP Port, SNMP, Powershell, WMI, SQL, HTTP, Perfmon, or Ping.
I do not check anything *nix, so I can not talk with that. But also for the Windows globe it is really straightforward to set up, exceptionally instinctive, and also exceptionally adaptable, It has actually really wonderful constructed - in control panel display screen, sms or e-mail alert, and so on
We've simply begun making use of "Servers Alive" which is really cost-effective, it isn't also rather looking, yet it sustains a tonne of various checks and also can signal in numerous means, takes care of professional scheduling/rosters etc for any kind of alerts. You can additionally make checks rely upon others, i.e. "this" system calls for "that" to be up/running.
We make use of WhatsUp from ipswitch, it is really simple for arrangement tiny networks, it can autodiscover networks by port check, it can make use of windows and also SNMP qualifications.
For display statics like cpu, mem, and also disk, we require to arrangement SNMP. WhatsUp assistance SNMP v1, v2, v3.
WhatsUp have an easy display via syslog (Unix), occasion visitor (Windows) and also SNMP Traps.
It has a wonderful ajax internet user interface with personalized customer and also personalized work spaces.
P.D. sorry for my negative english
I make use of a mix of Nagios, Cacti, personalized manuscripts and also among my very own projcts - - System Health Monitor. I such as having exterior solution monitoring along with charts of system sources so you can do upload - mortem evaluation of system troubles or promptly examine the charts to see if points look 'regular' contrasted to their historic values.
I've made use of hobbit, large bro and also nagios when benefiting poorer (read less costly) companies. Of the 3 I favor hobbit due to the fact that its straightforward and also bulletproof. I've constantly really felt that nagios is is attempting to be an open source variation of openview or tivoli, and also truthfully if I have the moment to invest setting up a structure like openview or tivoli after that monitoring is possibly my whole work and also my company can possibly manage to acquire openview, so why usage nagios?
Currently we make use of PRTG from Paessler. It is superb. No representatives called for, superb Ajax internet user interface, historic logging, graphing, WMI, etc etc There is a 10 sensing unit variation readily available absolutely free yet we stuck down a number of grand for the venture variation. Loan well invested.
For those that do not such as the Nagios internet user interface there is NPC, a plugin for Cacti that makes the Nagios UI readily available from within Cacti, yet with far better appearances (ajax etc).
It reviews from a data source given by NDO2DB, which is a wonderful means to have your framework readily available from within a data source for usage in manuscripts and also various other devices.
We make use of hyperic - it has both an open resource variation and also a business one
It checks the os (RHES 3, 4 and also 5+Ubuntu), Apache, MySql, JBoss, Tomcat, mail web servers, memcached and also it possibly can check extra applications. No unique arrangement is required, all web servers were located with the vehicle exploration, also if they were mounted in a renegade area. It is really simple to make use of and also set up, you can regulate your solutions (start/stop etc) and also specify signals.
Minuses - You require to configure it to work on boot (we are making use of cron to do that).
I'm making use of PA Server Monitor . It's largely Windows concentrated (occasion logs, efficiency counters, solutions, etc) although improving with various other systems since some minimal SNMP assistance has actually been included. Things I such as ideal is it's very easy to set up contrasted to a great deal of applications (no config documents, no command lines, etc). I would not advise it for a hefty *nix setting though.
Oh, it's not free, yet less costly than some rivals.
I've made use of Activexperts Network Monitor with wonderful success (on a primarily Windows network yet it had some unix and also linux hosts, printers of numerous brand names etc that was additionally checked with it).
It's actually very easy to arrangement and also find out, instead economical wherefore you get (was $500 for site/enterprise certificate) and also sustains vbscript and also remote unix commands. If the network is tiny (a couple of hundred nodes at the majority of) I assume this is far more instinctive than System Center Operations Manager which really feels extra routed at massive windows networks just.
Network Monitor features a great deal of predefined manuscripts for monitoring things like e - mail web servers consisting of numerous Exchange variations and also all its solutions, http web servers with predicted feedback, occasion logs, sql questions and also anticipated feedbacks and more ... and also dependences are very easy to set up (" all these rely on this router so if it falls short to reply to ping and also snmp, never mind startling us concerning all right stuff behind it that's not reacting"). SMS with portal or neighborhood GSM modem assistance and also all regulations can certainly have activities like solution reactivate, web server reactivate or personalized manuscript - to deal with reoccuring troubles for you (it's vital I assume, kinda like regression screening is for growth).
... I've additionally attempted to tame a Hobbit and also really did not actually appreciate it in all (neither the puffed up Windows representative) - yet it was set up for Windows web server monitoring and also it actually impacts at that - more than likely extra matched for a linux or unix - driven network.
I've made use of :
- Nagios - calls for some old - timey command line arrangement, not rather, yet tough and also useful. It has actually been superseded by :
- Zenoss - calls for a lot less maneuvering to set up, has a business version. As soon as running, the remainder is regulated via an internet browser. Really effective, yet calls for some MIB job if you make use of the free variation.
- Intermapper - business program, spendy if you have great deals of nodes to check. Seems created in Java (for far better or even worse).
- Spiceworks - have not attempted the most up to date variation. Older variations required a little bit even more umph under the hood to get it to react, yet or else, it functions perfectly. Free variation features nag advertisements.
PRTG Network Monitor - can not claim sufficient wonderful features of it. Outstanding internet front end and also specifically wonderful for monitoring routers (transmission capacity etc) and also various other tools via SNMP and also gauging uptime for SLA's, etc
I make use of a mix of Solarwinds, VMware web server performance tabs, and also personalized manuscripts.
Solarwinds Orion Network Performance Monitor is what I make use of with our Windows sys. admins on my internet servers. Still obtaining some valuable application metrics working on it, yet it has excellent details on standard box degree things (disk, network, CPU).
For my VMware visitors, I enjoy the performance tabs.
For my Sun web servers, when I require something that isn't readily available in Solarwinds (due to the fact that our admin hasn't included it or what), I write personalized manuscripts (generally in Perl) to check points like mirror wellness, swap use, etc
I would certainly such as to get even more onto Solarwinds, yet there's just like 26 hrs in a day (approximately my employer thinks) so I locate that can be a bit restricting ...
Smokeping not just checks the schedule of numerous web servers and also solutions yet additionally tracks their latency while giving very easy to make use of, wonderful looking, and also fast to present charts.
Variety of latency dimension plugins is readily available out of package. If you recognize some Perl, it is very easy to create your very own ones for any kind of unique demands.
Huge installments will certainly gain from Master/Slave System for dispersed dimension.
Very configurable signaling system will certainly aid you see concerns prior to they start influencing customers or advance right into significant blackout.
Smokeping is free and also OpenSource Software created in Perl by Tobi Oetiker, the maker of MRTG and also RRDtool
Zabbix. It's open - resource, and also sensibly straightforward to arrangement and also customise. We have a great deal of personalized monitoring manuscripts that feed right into the zabbix web server, yet it cares for centralising that information, presenting it suitably, alerts (email, IM, SMS, twitter, etc), etc.
I'm component of a functional monitoring upgrade task. We've had numerous suppliers come onsite to offer a couple of large buck systems and also combined in some less costly choices to contrast.
Among which is Hyperic, which is additionally readily available as a free open resource remedy. I was excited with its supplied capacities and also extensibility for personalized representatives.
Also have a look at Argent Guardian It's cross - system, can function as a syslog web server, they'll offer you the data source schema to do your very own coverage, if you require that, and also you can import your very own photos as "maps" to offer aesthetic signals.
Zenoss Core is of some usage, We are utilizing it (for concerning a year) for light-weight monitoring of web servers, net buttons and also UPSs.
Zenoss Core is an honor - winning open resource IT monitoring item that properly takes care of the arrangement, wellness and also efficiency of networks, web servers and also applications via a solitary, incorporated software.
For monitoring data (memory use, load, mysql task, apache task, etc) I make use of Munin. Out of package it currently tracks a great deal of points and also stories charts for various time periods (last 24 hrs, last 7 days, last month, in 2014). Via plugins a lot more points can be checked. It's result are HTML web pages with attractive charts.
Munin has a master/node style : nodes collect data on a web server and also the master shops the information and also generates HTML and also charts.
I make use of Monit to track running procedures and also to reactivate or signal me when particular configureable problems emerge (high cpu load, high memory use, no HTTP feedback, etc) Monit can additionally check even more basic features of a web server, such as cpu load, memory use, harddisk standing or disk use.
Monit requires to be set up for every single solution or equipment you intend to check and also just how to react when something fails. One of the most previously owned alternatives are to do absolutely nothing, send a sharp e-mail or reactivate the solution.
Monit is wonderful when it functions, yet occasionally it falls short to start, stop or reactivate a solution and also there is not a great deal of analysis details readily available to inform you what failed. This suggests you do not recognize if the trouble was with your solution or with the Monit arrangement, which keeps up a cron - like marginal setting.
Both tools are readily available by default on the majority of Linux circulations.
For the standing of web servers and also solutions (whether they are up or down, and also sending out cautions if they drop) and also for yes/no inquiries (" has a back-up been carried out in the last 24 hrs?") we make use of nagios. It is tough to set up, yet it is greatly configurable. Personalized manuscripts can be worked on remote computer systems. Alerts can send e-mails, send sms message or perhaps run personalized manuscripts.
For the wellness of web servers we make use of munin - it gives wonderful charts of memory use, cpu use, network use etc Pretty very easy to set up on linux at the very least (I have actually not attempted with Windows).
Nagios is wonderful given that it's free and also there is a lot of plugin's for it. Nonetheless the UI and also config is really hard.
It's specific reverse in pro's/ disadvantage's which is additionally wonderful is Microsoft System Centre Operations Manager (SCOM) which is not free, has much less plugin's yet arrangement and also config are great and also very easy.
I have to confess if I remained in a largely Microsoft firm, had really high dependence needs (i.e. can not manage for monitoring to damage) or needed to think of obtaining programmers to collaborate with it after that SCOM would certainly be my referral over Nagios.
It's vital for us to have really reduced expenses on each node to make sure that as several sources as feasible are readily available for calculation. Ganglia offers us an excellent review of the collection and also permits us to pierce to specific nodes if required. Besides recognize what's taking place now, we can get a respectable consider what's taken place over the last hr, day, week, month, and also year. The charts of numerous data are standard and also useful.
We've created our very own monitoring software. Our code isn't virtually as innovative as a business plan, yet we really did not require much capability. It was less complicated to write our very own than to explore various other plans and also find out just how to utilize them. The code does simply what we desire and also it's very easy to expand.
We make use of (and also like) WhatsUp from Ipswitch for our reasonably tiny Windows network. It is very easy to arrangement, and also reasonably very easy to take care of, and also recognizes just how to manage Windows web servers along with typical things.
For bigger networks, non-Windows-oriented networks, or connect with great deals of diverse things, I exuberantly advise OpenNMS. OpenNMS software program if free and also the firm is greater than satisfied to market assistance and also execution solutions. It additionally takes place to be run by a really sharp friend of mine from university!
I have actually been doing turn out of Spiceworks at our firm and also we are locating it to be a wonderful device not simply for monitoring web servers yet every little thing else on the network.
It does points like automated supply and also personalized monitoring to send you e-mails when there is a trouble (EG : Printer is to 10% of ink or disk drive of this web server has 20%).
Its disadvantage would possibly be is thickness of details per computer system, do not get it incorrect it has A LOT of information per equipment but also for points like web servers where you could desire a great deal of statistics you could require to make use of an additional device.
EDIT : oh did i state its organisation version is based around it being free for life.
Personally, I enjoy Munin which is really simple to install and also to write plugins for as it has a really uncomplicated style. There are fairly several plugins currently around for all the objectives you can visualize, so you possibly will not also need to write plugins to begin with.
It additionally gives attractive charts and also the alternative to set up (really standard ) signals.
Cacti is a great online frontend to RRDTool, giving really convenient charts and also statistics. RRDTool is the component that collects information from numerous systems and also checks a variety of technological information.
We're making use of that cacti/RRDTool remedy to check Unix and also Windows systems. We get a great deal of valuable metrics consisting of load, CPU/RAM use, HD room, customers visited, network website traffic, running procedures, and more.
You will certainly find even more details on cacti on the What is Cacti? web page.