LogicMonitor Founder On Exposing Amazon Performance Issues, Fixing Cloud Sprawl, And Making MSPs The Heroes

'We Allow The MSP To Become The Hero'

Steve Francis, founder of the fast-growing SaaS-based performance monitoring vendor LogicMonitor, spoke with CRN about a wide range of issues including Amazon Web Services' performance problems, cloud and monitoring sprawl, and how his platform can save enterprises millions.

Francis ran the network and data center operations for National Geographic Society and the University of California before founding LogicMonitor a decade ago, which now monitors technologies from AWS, Cisco Systems, NetApp, Palo Alto Networks and VMware.

With the ability to monitor both legacy and new technologies either on-premise or in the cloud through a single platform, Francis said LogicMonitor is upping the game for what solution providers can do for their customers.

How are your MSPs exposing and fixing AWS performance holes?

AWS recently had an issue that we saw in our internal system because we use part of Amazon for some of our product delivery, but we run our own servers on our data center. … At this particular time, we were getting calls that reports were slow to download. But because we have all this monitoring, we can check the time to query Amazon S3 resources not only by using CloudWatch -- which Amazon tells you everything is great -- but by actually checking the real response time from different locations around the internet and from the customer's office locations. … So you can quickly identify where the issue is.

So how early on did you know about the performance issue before it was reported by Amazon?

Pretty much immediately. Sometimes we see issues that Amazon never acknowledges. If it's fairly short, maybe two hours, we'll see that Amazon had a performance issue. We can tell it was from all these locations, so it's not just one ISP or anything, and Amazon doesn't acknowledge it. We monitor not only what you have in the cloud, but we also monitor the actual providers. We monitor AWS as well to let you know it's OK. It's a more holistic approach versus other monitoring tools.

Did Amazon come clean on their performance issue?

No they didn't. The recent performance issue, they didn’t.

Talk about how big a deal it is that LogicMonitor can expose Amazon issues, and explain last week's issue.

It's unclear how big an issue this particular one was because it was intermittent and it was mostly out of their LA data center. We saw it because of our own reports, our own infrastructure and we monitor our own stuff very well. So there would be cases where the reports would take minutes to download and that existed for 10 or 15 minutes, and then it would clear up and be fine for 10 minutes, and then it would take minutes again. Usually those reports are instant. [The reports] were delayed long enough that because you're looking at these reports in a browser, they took long enough that most people thought, 'Oh, it's just broken and I'm not waiting anymore.' A giving-up point of view. Basically, functionally down.

How long was it down for?

About an hour of total time, which is short enough that Amazon gets away with it.

So how important is it for a solution provider to have LogicMonitor's platform?

If you're the VAR or MSP that is setting up services for your enterprise customer and your enterprise customer is coming to you saying, 'My performance sucks. Why aren't my reports or other issues being delivered?' And you just look at Amazon and say, 'I don't know. There's no issue.' They're going to nail you to the wall because they have no visibility into Amazon and Amazon doesn't mention it. But if you have a tool that says, 'I'll set this up for you. I'm also monitoring it for you from multiple locations. I can show you right here that Amazon had this issue, it wasn't just us, and it was from this location.' Then you have accountability. Then you have accountability to go to your Amazon account and say, 'What the heck?'

How important is it for solution providers to monitor in a multi-cloud era?

If the partners are being engaged for cloud transformation, which they are, and they're recommending any form of public cloud, then when that public cloud has an issue -- who do you think the enterprise is going to talk to? It's going to be the MSP. You need a tool that says, 'This is the issue. Our infrastructure works correctly. We had a design for failover.' But in many cases MSPs are always going to be held accountable for the availability and performance of any solution they implement, even if it's a cloud solution. So they need the visibility. … We allow the MSP to become the hero.

Can you give an example of how an MSP can become a hero?

We just did an internal project with our own technical operations team where we looked at AWS and decided, 'How much can we save here because of how we monitor it?' We saved probably $1 million a year because of that.

Explain how you saved $1 million per year.

With our monitoring, you can look at the monitoring of all of your AWS resources and say, 'Show me stuff that is underutilized.' So you find 15 [Amazon EC2 instances] that have had zero CPU utilization for the last two months. So why are we paying for extra-large mediums for these machines that don’t do anything? That's one aspect, just identification.

That's a real problem if you try to do it through Amazon's dashboards. If you want to say show me my least utilized resources, you have to go in region by region by region -- and you can't look at it all at once. With LogicMonitor, you can say, 'Show all my resources, no matter where they are, by utilization. Show me the least used 25.' Then you can figure out what you don’t need anymore. … You think virtual machine sprawl was bad, wait until you get cloud sprawl.

How bad is cloud sprawl right now?

It's very easy to just blow out your budget without even knowing why. Because there's no cost to bring something up effectively … there's no system to say, 'I only want to bring this up for two hours then shut it down.' So they'll bring it up and forget about it or bring it up for a week to do a project, then get distracted afterwards. Then all of a sudden you end up with thousands of machines running just doing nothing. That's an easy situation to get into.

How much monitoring sprawl are you seeing?

Enterprises often grow through acquisitions, so they'll have 30, 40 up to 100 different monitoring systems. For example, one business unit started with networking, then they bought another business unit and they used something else and never changed. Then another one brought SolarWinds on, and another one brought ScienceLogic and they just run them all. You end up with a lot of overhead of people, hardware, expensive data center space that is dedicated to racks and machines that do nothing but monitoring, then disaster recovery – it gets very expensive.

What's the biggest myth about LogicMonitor in terms of what you can do?

The biggest myth is that we don’t exist. Awareness is probably our biggest problem. If you look at what we do, monitoring isn't a new issue, lots of companies have done monitoring. What we do differently is we deliver through SaaS, which means it's super easy to deploy and roll out. That's a less strong sounding differentiator except if you’re an MSP because ideally you're deploying to a new customer every week. So the speed of deployment really matters.

How important is that SaaS speed of deployment for MSPs in the enterprise?

If you’re an enterprise, you're not deploying all the time, but you're changing your technology fairly frequently – rolling our new services, building new applications – so the speed of deployment really, really matters. We do it through SaaS to make the deployment easy and we have the visibility of your old, legacy stuff – the NetApp and Oracle running on AIX -- through to the cloud solutions and providing that whole visibility at once.

So if you're doing any kind of modernization – moving stuff here to there – you can't be looking at two separate monitoring systems because if your new stuff says, 'I'm performing slow' -- is that's because it's slow or because it's talking to a database that is slow? You need that in one pool. Otherwise, your time to repair is going to be much longer.

What separates LogicMonitor from your SaaS-based monitoring competitor DataDog?

One difference is that they're agent-based. So in order for DataDog to monitor resources, they want you to put their code on your servers. We are agentless, so we don't have that code distribution and updating problem. The other big difference is they don’t have the ability to look at the traditional data center and say, 'Monitor the Cisco gear, the Juniper gear, the NetApp, all those storage arrays, and so forth.' The big differentiator for us is, if you're running in the data center and in the cloud, we're a far better fit. If you're running purely in the cloud, depending on your point of view, some may prefer them.

What separates LogicMonitor from your on-premise competitor ScienceLogic?

We do run into ScienceLogic a lot in the service provider space. We usually win against them. The main differentiator is we have a truly multitenant SaaS model. Then it's also just the automation and ease of use that we have. Their user interface is less elegant. They don't have a multitenant SaaS model and they are on-premise. They can't go into the cloud.

Does anybody else have both the on-premise and cloud capabilities like LogicMonitor?

No. Not delivered through a SaaS model.

Where do you see the channel opportunities with LogicMonitor?

Transformation opportunities are a huge boon for the channel. You can't read any magazine targetingt a CIO without reading about IT transformation or digital transformation – whatever you want to call it. All the CIOs are thinking about it and know they need help with it because their internal staff is focused on keeping the lights running on all their data center equipment and they don't have the forward-thinking technology that can do this. This is a huge opportunity for MSPs and solution providers.

What type of solution provider is coming to you?

I'm don't know the exact breakdown, but we have solution providers that are purely networking guys. We have solution providers that were VARs that want to get into managed services. Also big companies like Teklinks that do some of everything.

What are you seeing in terms of the channel growth?

We'll be growing 90 percent year over year in revenue this year. More growth is always better, so if the channel decided they wanted to move faster with adoption with us, that would be great.

We have good training programs that help people get the value out of LogicMonitor. We also do training on what services opportunities are available, even understanding what services are available in Amazon, on Azure and Cloud Foundry – there's so many of them.

Is the channel moving fast enough here?

It's going pretty well on the channel side. They're becoming pretty progressive in deciding that this is the kind of toolset they need. There are some providers that are saying, 'IT transformation is the wedge into enterprises and we need to get ahead of that and adopt it.' Those channel partners are seeing really good success. A lot of them don't seem to have discovered that yet. If you talk to any CIO, that is what they're talking about and they know they don’t have the resources in-house. They are looking for the help to do it in a lot of cases. We are offering a huge opportunity in the channel. They really should be faster and more proactive about positioning themselves around it.