What needs improvement with Device42?

The product must provide AI features. It would be very useful if I could create datasets or queries from an AI interface.

In my experience, I believe that the key concern is the pricing strategy of the solution. Instead, other solutions are much more cost-effective. Previously, Device42 has altered the pricing model to include a subscription fee which I see as very costly. They should change their pricing strategy and license scheme, conduct market research and ensure that they provide the right product in the market at the right price.

I would like to see API management as an additional feature in the tool's future versions. It will give more API security.

Mapping items wasn't as intuitive as importing in Device42, so this is an area for improvement. The real-time configuration also has room for improvement in the solution. Importing was easy, but once you had all your data there, it was hard to find precisely where everything was except for basic information. For example, when you go to the program list of what's on the Linux box, as my company was using Linux boxes, it was a little harder to find that information, and it was also hard to recategorize. If you imported your data wrong, you had to start over, and you couldn't just drag and drop items to other places. This limitation, I feel, is due to Device42 wanting you to ensure you have the correct configuration when you import so that other people in your organization can't just move stuff around. I want Device42 to make it easier to import and move items around and then add a lock feature, so people won't be able to do anything to the items after. I also want better integration between Device42 and Jira because my company uses Jira as a base to make sure nothing's missing. Jira has been in the company even before Device42, so my company imported all data into Jira and then used Device42 for mapping. Device42 is housed in a database, while my company used Jira for the front-end view of all items. The issue with the integration between Jira and Device42 is that Jira has a language that isn't compatible with Device42, so integration with the two should be improved, or at least there should be a mapping function where you can tell it what to display on Jira and Device42.

Configuring rPDUs in the data center for Device42 required us to add multiple discovery jobs. The rPDUs were referred to the ISPDUs and daisy-chained. They would automatically assign a unique SMP port, but Device42 didn't allow SMP ranges, so we had to create a discovery job for each rPDU individually. We submitted a feature request to mitigate that.

The architecture is a bit old-fashioned. Device42 is on one server and runs as a closed appliance as a virtual machine or guest. We are loading more into Device42 than it can hold. Overloading Device42 with REST API calls or tasks will directly impact every aspect because the server will be too busy to answer requests. that is why we invented some creative ways to reduce that load from devicd42. A better system would be a frontend server with a backend server and a database-like architecture. The fact that everything is one server causes a significant performance problem. It would be better if Device42 could restructure the product and make it more scalable.

It was hard to know which assets I'd already looked at because if I looked up another asset with a similar name, they would still pop up. I would have to make sure I was looking at the right asset. The search function should be improved.

The reporting could be better. The Insights+ component is an improvement, but they even admit it isn't the greatest. The documentation on their website could also be improved, but Device42 is constantly changing, so the documentation would have to change almost daily.

The breadth and depth of the solution's discovery of IT asset information means the system does most of the work, but they need to work on improving the database discovery part, especially for Oracle Exadata. Also, the analytics reporting is very good, but there is a little bit missing when it comes to documentation about custom queries. Another area for improvement is being able to distinguish the services on the individual machines. There's a lot of data and you can't map it. We need to clear the data. Because there is so much, we're not able to get what we want.

They always say and recommend that we should run one remote connector, and one discovery service for about 500 to 1,000 devices, which actually means the productivity of this remote connector. Even then, probably in different subnets, they'll also advise you to have different connectors. If they could make it so their remote connector could do as much as 10,000 devices, that would be better. Just one remote connector to manage about 10,000 devices and still not reduce its productivity would be perfect. That would be a bigger picture item. Then, we wouldn't need to install or set up more than one remote connector on different subnets. That's one thing they could do to improve their services. It would be nice if they could add support tickets into IT service management.

A con for Device42 is that Kubernetes integration is lacking. You pay for 10,000 spot licenses and if you're spinning up a Kubernetes cluster, or four or five or six Kubernetes clusters like we do, you're going to have 5,000 or 6,000 nodes in each of those, doing different types of business things. And those don't flow in and out. Yes, you can purge them — they do have a purge function — but the way that that works for security purposes leaves a little bit to be desired. I want to keep track of those for a security solution. If there were a compromise, it would be nice to have that rather than have it taking up some of your spots. Something that's missing is Kubernetes-based statistics so you can get a peek inside your clusters there. Also, the reports are a little clunky, they could be better. Finally, the comprehensiveness of the asset discovery is not perfect, but that's part of it being open-source. The scalability, the time that it takes to get things inside, could be faster.

I tried in the beginning to use the agentless discovery. I had a few problems. There was an employee before me who started with Device42 in 2018. I came to the company in Summer 2019 and was the new tool owner. We had some problems with the agentless, so we switched to the agent discovery. We have a big Linux environment with a lot of Linux services. All Linux services are orchestrated. Therefore, I had no problems rolling out the agent approach to 2,000 to 3,000 machines. The scheduled task for the control for the agent I can create this in a few minutes with the orchestration. I have a feature requests for the UI to sort numbers or estimates. When servers have two network adapters, automatically discovered will be only one network card because the other one is a backup. Device42 has some problems to find the other connection. I have open some tickets for this. For some of the network devices, the auto-discovery could be a bit improved. Some switchboards for some devices have problems to receive their connections automatically, and support names are sometimes incorrect. These are only for special network devices when we have a clustered network device.

The solution’s automatic IT asset discovery and inventory functionality functions pretty well. There could be some improvement if there were some automated scripts to get it off the ground. I know it takes a bit of effort to get all of various managed devices into one place. We have to go and change how they are managed and make sure they are all linking up correctly so they can be tracked in Device42. Some type of automated script for each to get over that initial activation. Once everything is in and tracked, it is a very robust solution. By the end of this year, I am looking forward to getting it rolled out and properly set up in our organization. It gets a bit confusing sometimes when you're trying to map things to a certain spot when you are trying to configure it. The only trouble we have had with it is the initial activation to get started and get things linked in there.

Device42 is a main part of our processes. We need reliability, not only in terms of the data but with the solution itself. It's really difficult when we have 10 minutes of Device42 downtime because none of our teams can work for those 10 minutes, and it's more time lost if there is longer downtime. An improvement would be to have a cluster implementation of Device42 to have high-availability and ensure that we don't have downtime in case of failure. Also, when we do an upgrade, it's mandatory to shut down the application. It takes 10 or 15 minutes to upgrade and it's once a month. It's not a big deal, but if it were possible to have no downtime, that would be better. We can plan for upgrades, but we try to have 24/7 availability in our organization, so there really isn't a good time to shut down service. Doing so always has an impact. We have teams around the world.

Sometimes we need to add some details on the purchasing side of the tool. If we have a contract, we need to keep track of the beginning of the contract, the end date of the contract, and a notification period when we need to decide whether we are going to cancel it or renew it. This is related to the reporting because we wanted to add these details in the list which we see on the tool, but even when we succeeded with that, we couldn't see this information on a report. But overall it's an easy program and we can manage it without too much trouble.

The overall speed could be improved. Sometimes it's very slow to do certain functions. General navigation is quick but if I want to delete an asset from a cabinet it does take a while. And if I'm doing it in bulk — say, for example, if we have one cabinet that has 20 servers in — if I want to remove all 20 servers, I have to do them individually, which is a bit time-consuming. If there were a way that I could just bulk-remove everything from there, that would definitely save some time. Also, I know there is a function where you can add cables in Device42, as in you can say: So-and-so port on this server is connected to this port on that server, for example. But that's quite a drawn-out process, trying to use the spreadsheet just to add a single cable. I think that could be simplified a lot. You could just have a column for the device name, the port, and the end-device port. If it could be populated that way it would be a lot easier. At the moment, the spreadsheet that you have to use to add cables has a whole bunch of columns which I think are irrelevant and shouldn't be there.

My biggest problem with the product is the upgrades. First, we have to do them manually and second, not this last time but the time before that, we actually had to build a new VM to deploy the solution again. We had to back it up and then restore it to the new version. That was inconvenient. I would also like the ability to bulk-delete. The application is great, for the most part. There are just a few little bugs and quirks that are annoying to me.

Since I was focused on deploying connectors and getting all the servers to be scanned, one of the biggest pains was when a job would fail, then the output (logging) was poor. For example, "Why did it fail?" In these cases, you get a generic error. It doesn't point you in the right direction and tell you why you got the error, which is really annoying. There have been times I asked, "Is there somewhere I can see a better log as to why is this failing?" That would be a really nice improvement.

While the automatic IT asset discovery is great, the first time using it can be confusing when you are configuring the SNMP. I don't remember for sure but I don't think it said "SNMP community," it said "password". The first time I used it I was thinking about communities but the tool said "password," and when you say "password" in SNMP you are thinking about SNMP version 3. This is the only thing that is confusing, although there were some devices that were not included in the discovery.

For the iPhone, the device diagram cannot be downloaded. You have to go to each device and download the device diagram for that device. E.g., suppose there are a 1000 or 2000 servers. You cannot go to each server or device to download a device diagram for each device. There should be a one-click solution for downloading all device diagrams.

Room for improvement would be in the discovery. Although the discovery it does is really good, there are certain elements that could be better in terms of a deeper discovery. An example would be teaming on Windows Server. It doesn't currently pick up that functionality well. Also, the dependency mapping can be quite slow sometimes, if you've got a lot of things connecting to services. It can be very slow to build up the map. The certificate management could also be a little bit better. Finally, it would be good to introduce a mobile app. At the moment, you can connect to the web interface through your mobile and select a mobile view, but it's still very much a desktop view on your mobile. It's not very mobile-friendly. So it would be good either to have a mobile app, or a specific mobile endpoint on the web front. You would visit a separate URL or it would detect that you're on a mobile and rearrange the view to be mobile-friendly, like dynamic websites do at the moment.

It would be nice for the agent to have an installer versus a single file across multiple systems. It'd be nice to have the reporting be a bit more user-friendly. The workflow could be a bit more intuitive. Though, the support has been great on helping us when we have ran into educational issues.

One thing I have brought up with their product team, and we opened up a case and a feature request for it, is having a requisition-style system in the software. Right now it does have the ability to track purchases, which is okay. But I would like the ability to create a request and say, for example, I want a new server. The way that it works right now is that you have to create the server object before you can assign it to an actual purchase. Obviously you won't have the server until you've bought it, so it's something of a cart-before-the-horse scenario where, in order for me to create a purchase, I already have to have the device, which doesn't make sense. It's fine if you're entering the data after the fact, but I would hope to use the system in a pre-purchase scenario, where I or one of my technicians could log into the system, create the request, and then forward that request to my purchasing people or to our ERP. The other area where it causes us a little bit of a hassle is if I want to track the cost of spare parts. Recently I had to order about a dozen hard drives for servers. For me to use Device42, I had to wait until the drives came in so that I could create the drives and assign the serial numbers to them, and then assign a cost to each drive. A better scenario would be where I'd say, "Hey, I'm buying 12 drives," and be able to have 12 generic drives in the system without serial numbers. Then, once the drives came in we could assign serial numbers and then further assign drives to servers as we need to. I worked with their product team at least a couple of years ago to express that concern. They claimed to like the idea, but I haven't seen it implemented yet. I don't know if it will ever make the cut. And I would like it not to be a paid upgrade feature. If they could include it with the license we use, that would be extra nice.