What is our primary use case?
We use it in a traditional way. We replaced an older system based on Nagios, and we had problems with false positives. We install the Zabbix agent on the client hosts and set up automatic installs. The agent is added to these other client hosts, and we have the infrastructure in place to support static servers.
Then we write the rules to tell Zabbix how we want it to report, and then let it run. If we've got a critical error, someone will be notified. Otherwise, we get an email message that a server is having a problem.
Sometimes, the problems resolve themselves, and sometimes, they don't, and we have to take action. But it's mostly hands-off for the most part. It's doing very well for that.
How has it helped my organization?
It has improved our overall server performance monitoring. We know right away when there are problems. It has built-in statistics, so we can identify if there's spiking. We can see what's happening every day around the same time and check the configuration to see if there's something that's malfunctioning and needs to be fixed. As we get better at diagnosis, we can go back and look at the history of the systems to pinpoint problems.
What is most valuable?
I can pull up graphs using the parameters I want to check. You can see the current day, week, or month. It's all built-in. We use it to bring up graphs of specific parameters.
What needs improvement?
Sometimes, the documentation is a little bit awkwardly written. The language barrier and translation to English can sometimes make it difficult to understand what they're trying to get at. It's just a language thing.
We can usually figure it out, but that's the biggest thing I can think of. Other than that, it's been extremely reliable.
For how long have I used the solution?
We have been using it for four years now.
What do I think about the stability of the solution?
I would rate the solution a nine out of ten; it's still really stable. We rarely have problems. And it's always been something else, like the database acting up or whatever. It's usually an external thing. It's usually not the Zabbix code itself.
What do I think about the scalability of the solution?
Scalability is good. Zabbix uses proxies to manage incoming data from hosts. Since the system is receiving a lot of data, that could potentially put a real load on the server. The proxies handle some of the load from their respective clients and then feed that to the main server.
For checks, Zabbix prefers that you use checks based on the data that's automatically flowing from each host rather than doing explicit checks or things like... don't go on this server and run a command except for a check. Just use the built-in checks, and we mostly do that.
Some people when they switch over to Zabbix, they're not used to that. They're used to always adding regular checks, but sometimes it won't work. So, you can change it to be more on the passive check side, where the hosts don't have to do anything, and it's actually displayed on the server.
How are customer service and support?
We've almost never used their direct customer service. We use their forums if we have issues that aren't super critical but still need addressing. It's mostly community support, which is free, and we can get a bit of information by asking questions there. It's not really direct support, but more like community support.
We actually paid for support for the last couple of years and decided that we've got it under control. We're not paying for support anymore, so we feel confident enough to manage it ourselves.
How would you rate customer service and support?
Which solution did I use previously and why did I switch?
The previous product was really outdated and, literally, it was broken. It should have been updated years earlier.
It was running but struggling. For example, it would get behind on its checks, and once it got too far behind, it would just throw them all away and start over. It was constantly missing things. It was clearly overloaded and not designed for our scale, so we were happy to move away from it. The difference after switching to Zabbix has been night and day.
I used to get paged at least once a day, sometimes multiple times. Now, with Zabbix, there might be weeks without any pages. We've never had a problem where we missed an alert.
It's great that the checks are much more intelligent. For example, rather than just saying, tell me when the load goes over five, you can say, "Tell me if the load is over five for more than ten minutes." So, if it doesn't go over five for ten minutes, you'll never get notified. It only waits for a combination of conditions to be true before it alerts you. That really cuts down a lot on the false positives.
We used a heavily customized Nagios. I wouldn't compare this old version to whatever is current so this is not a criticism of Nagios.
What about the implementation team?
I'm not the person who set it up. We have a person in charge; he did most of the work. And then a lot of us – most of us do our tech work on the templates and things. We don't really work on the core of the product.
It's a small team. One main person and then two or three support people help them with specifics, like setting up the databases and things like that. So, supporting those kind of tasks. So, it's maybe three or four people. And many of us know how to use most of the features. Some people are better at things in other situations, so we just need to fully share that knowledge.
The main person's role is systems administrator.
We have one to two people who perform maintenance. The person who runs the updates isn't the main developer. We have another admin who is really good at that. We'll all connect to a remote call like a Zoom call and watch the update process. This lets us share information about how the system is working.
What's my experience with pricing, setup cost, and licensing?
For pricing, it's free. We don't pay anything for it. They open-source the code, and people pay for support. We purchased support for a year but never used it. We dropped support again since we have a tight budget.
It's been very reliable.
What other advice do I have?
I would recommend Zabbix to others. We've been very pleased with how well it's worked and how much of an improvement it is on the system we had before. It's night and day.
I would rate it a nine out of ten. Nothing is perfect. One thing is a little clunky about it. We provide this hosting service to a bunch of different groups, and by default, you can only see the servers that are in our group. So, when we log in with our normal IDs, all we can see are the servers from our monitoring group. If I want to see other servers, I have to log in differently. (Turns out I had a parameter set incorrectly, and this problem is solved now)
Which deployment model are you using for this solution?
On-premises
