Kentik Reviews

For our purposes, where we're at today, and even in the past, to analyze flows and to pull specific data and understand where our traffic is going to — which AS path — that's primarily the value that I extrapolate from Kentik.

It's mostly on-prem. We do some stuff with GCP and AWS, but it was all primarily licensed-based, based on the number of pieces of equipment we have on-prem that we actually attach it to. We have over 55 edge nodes and about 10 compute nodes.

We can actually see what we're doing now. When it comes to making an educated decision on a number of things, if you have no visibility into what you're doing, you really can't make that decision. Collecting that data and having those metrics first-hand, in real-time, allows us to make an educated decision, versus an uneducated guess.

Kentik has proactively detected network performance degradation, availability issues, and anomalies. When we had no visibility. When we had congestion, things would actually happen and it was hard to troubleshoot as to where they were coming from. That was one of the first things we were able to do. 

A specific example is where we had a number of tenants that were created that were getting DDoS'ed. We couldn't understand how or why we were getting DDoS'ed because we had no visibility. We were guessing. Kentik opened up and showed us where the traffic was coming from and how we could go about mitigating.

It lets us understand what those attacks are, versus not actually knowing where they're coming from or how they're affecting us. It cuts down the time it takes for us to troubleshoot and actually mitigate by about 50 percent, guaranteed, if not more. But we're running a bunch of GRE IP sectionals. It's not like we have huge amounts of capacity. But for some of our large customers, it really has helped us detect what the problem is, instead of guessing.

At my previous company, it improved our total network uptime by about 20 percent. I wouldn't correlate that back to Kentik in my current company.

The most valuable feature is being able to pull traffic patterns; to and from destinations. We're able to understand where our traffic is going, our top talkers from an AS set, as well as where our traffic's coming from.

The only downside to Kentik, something that I don't like, is that it's great that it shows you where these anomalies lie, but it's not actionable. Kentik is valuable, don't get me wrong, but if it had an actionable piece to it... I keep telling them, "Man, you need to find a way to make it actionable because if you could actually mitigate, it'd be huge what you guys could do."

The way things are, we have to have some sort of DDoS mitigation, like Arbor or something of that nature. Once the anomaly is detected, that's great, but then you have to mitigate. If Kentik had mitigation, or if they could acquire a solution and throw it onto their platform and have that portion available, that would be huge.

I have been using Kentik at this company for about a year and, prior to that, I used it a previous job for about another year.

Coming into this company, I felt they were flying blind, meaning they didn't really have anything from a monitoring standpoint. They didn't understand how decisions were made. And to make educated decisions, you actually have to have the proper tools in place. Kentik was a tool that I know works really well.

Kentik has pretty good intuition, as a company, as to where the market sits and what they're into. They don't delude themselves. They really focus. They've been pretty good. I know the leadership over there and it seems like between Justin and Avi, they're good at what they do and that's why I'll continue to use them.

Anywhere I go, I'm going to use Kentik if I have the chance.

We use it almost exclusively for flow data. We use that for a variety of things from network optimization to network capacity to security events, including DDoS protection, etc.

We're using the SaaS version.

The drill-down into detailed views of network activity helps us to quickly pinpoint locations and causes. Anecdotally, it has decreased our mean time to remediation. On a per-incident basis, it could save anywhere from five minutes to 60 minutes.

We also believe it has improved our total network uptime. We haven't done any direct before-and-after comparison, though.

Again, anecdotally, it has sped up our security team's ability to respond to attacks that did not surface as readily, prior to having the flow log data.

One of the valuable features is the intuitive nature of building out reports, and then triggering actions based on specific metrics from those reports. It has a really good UI and the ability to surface data through the reporting functions is pretty good. That's helped a lot in the security space. If you get a massive, 100 GB attack coming through, saturating links, you can surface that really quickly and then act to engage DDoS protection or other mitigations from the IPS.

The real-time visibility across our network infrastructure is really good. One of the things that we love it for is our global backbone visualization. Being able to see that utilization in real-time is pretty critical for us.

It also proactively detects network performance degradation and things like availability issues and anomalies when used in concert with the SevOne network management system. In conjunction with that — with all of our polling and availability data coming from that NMS — the flow data provides that type of insight.

We also use Kentik's months of historical data for forensic work. We do 90 days.

I believe they're already working on this, but I would love for them to create better integrations from network flow data to application performance — tracing — so that we could overlay that data more readily. With more companies going hybrid, flow logs and flow data, whether it be VPC or on-prem, matched with application performance and trace data, is pretty important.

The other area would be supplanting companies like SevOne and other companies that are really good in the NMS space, specifically for SNMP data.

We've had it since before I took over this space and took over Kentik, so 2017 is when the initial contract started. We're going on three years.

The stability has been very good. There was only one outage or impacting event that I can remember in the past year. It took them a couple of days to fix it, but the impact was remediated through some mitigation they did on their end to prevent it from causing us too much headache. They got it down to where it only affected some long-term reporting, which wasn't super-critical for us. It wasn't too big a deal.

So far, Kentik has scaled for what we've done with it and we haven't hit any scale issues to date. I don't know if we're a very large user compared to some of their other customers so I don't know if we're a good example to discuss scale, per se. But we haven't encountered any scale issues from our side.

We don't have plans to expand the use of Kentik, other than increasing licenses to gather flow data for more devices. We buy per license and we have 75 or 100 licenses. The size of the teams that use it is 100 people or so. They are security engineers, network engineers, network health analysts, and threat-intelligence folks.

Their tech support is phenomenal. They tell us about an issue before we even get to it.

With the incident that I mentioned in the context of the solution's stability, even before we experienced any issues relating to it, they had already reached out to us and let us know what was going on. They gave us some timelines, and the ongoing communication kept us informed throughout the incident and was able to mitigate any kerfuffle from the executive layer. That can be a giant headache when dealing with those types of situations, but they managed it perfectly and were proactive with their communication and we didn't hear a peep from anyone about it.

I wasn't involved in the initial setup, but there is time involved for us to set up the checks for the flow data and to set up the reports. Depending on what someone is setting up, it could take five minutes or it could take a couple of days. It just depends on what they're implementing with it.

I'm sure we have data available to show ROI but I don't have it available. Where Kentik is bringing us the most value is in the security realm, in terms of attack prevention, but ROI on that is hard to measure.

There have been other folks in our company who have tested a variety of things. Prior to Kentik they went through an evaluation phase, from what I understand, and vetted out a variety of solutions. I believe that what made Kentik stand out was pricing and the intuitive user-experience.

The biggest lesson in using Kentik is that as we continue to use it and learn more, we learn about the use cases that are valuable. Initially, when I came over to the team, we weren't using it to its fullest capabilities. As we started to understand the capabilities and dive in, in specific areas with Kentik engineers themselves for customer success, we learned that we needed to change our thought process a little bit; how we thought about flow logs and what they could provide insight into.

My advice would be to leverage their customer success engineers upfront and don't let them go until you've hit all your use cases. Constantly be in touch with them to understand what some of the forward-thinking ideas are and what some of the cutting-edge use cases are that their other customers might be getting into.

We don't make use of Kentik's ability to overlay multiple datasets, like orchestration, public cloud infrastructure, network paths, or threat data onto our existing data. That is something we're evaluating. We're currently talking with a couple of teams that are moving to AWS, teams that would like to use Kentik to potentially capture VPC flow logs and overlay that with their application performance data. That is something that is currently on-hold, pending some other priority work. We will probably dive back into that, with that team, around mid-2020.

For maintenance, it requires less than one full-time engineer because it's a SaaS model.

In terms of overall vendor partnership, I'd give Kentik a nine out of 10. They're right up there as one of my best partners to work with, amongst all the contracts that I own. They're very customer-centric. They're always available. There's nothing too small or too big that I can't ask them to help with, and they seem to be willing and able to jump in no matter what. That customer focus — which is a theme across the digital world right now with companies trying to try to do more of that — Kentik does a really good job of embodying that.

The DDoS alerting was, at first, the most useful. It was able to alert the entire team of more than 20 that the issues with the website were actually network based, instead of, say, bad code. In time, we mitigated the DDoS attack surface, so the usefulness is still there. We just don't see it every day.

Now we use Kentik for more nuanced traffic insight. This is ad hoc usually, but we do email 'peering' reports daily to the lead network engineers. This gives them some view into new traffic patterns we are picking up in IXes.

I find it very useful to see when traffic destined for a prefix that we prefer ingress on in the East Coast actually ingresses or egresses on the West Coast. It shows the difference between BGP paths vs. regional expectations.

The alerting ability is greatly improved. I think there is some movement still to make this into a 'dumb mode' vs 'expert mode'. There is the SQL-like syntax, but that is expert+.

I have used Kentik for 2.5 years.

We rarely, if ever, had any stability issues.

I have not had any scalability issues.

Technical support is second to none.

We used in-house, hand-built things. All based on binary RRDs or worse.

Initial setup was very straightforward. Nothing I needed too much help with.

There is a large difference between BGP and normal nodes. I don't think this plays out to the best for the customer or Kentik. To be able to split off the BGP vs PPS requirements would be good.

We've evaluated almost everything except SiLK.

Use the technical support if you need it. They are excellent.

• Dashboard visualization: I have come to rely on Kentik for a constantly refreshing view that I can quickly customize. There are a lot of graph options, including a flow diagram that shows how traffic has moved over our platform.

• Kentik Detect: The Detect section of this product provides an excellent way to utilize flow in order to produce alerts. It is very flexible and allows for a number of interesting customizations.

• Sankey flow diagram: It shows the path that the data takes through my network and in turn allowing me to troubleshoot many issues, without running through a crazy amount of device configs.

• SQL query: You can directly query the database that Kentik keeps and, via API calls, make it something you can easily integrate into your own product or monitoring tools, or just to produce reports with metrics that might not be a Kentik feature yet.

We have put it on half of our large monitoring screens. Sometimes, it is actually easier to identify and attack incoming traffic using Kentik, than it is to use our own gear.

Even when we know what the traffic is, it allows us to jump directly into the next steps of our process more quickly, since we can visually see everything in one place and on one screen through the customizable dashboards.

Instead of just total traffic in bits or packets, we can get protocol, destination port, TCP flags; everything you might want.

Kentik has been remarkable at anticipating the design requirements of their customers. They have provided everything that I might want already. After using it for over six months constantly, I am still discovering new things.

The only times I’ve felt that “I wish I could use this to XYZ,” I’ve contacted support and it turns out that I can do that already. However, I just didn’t know if I could do it using the existing controls or via a combination of query types.

Perhaps a better explanation would be to see how tagging is captured and a method of comparing my tagged interfaces on Kentik’s side. Right now, I can go in and look at all of the interfaces that they’re receiving the flow for and also sort/filter it, but there is no way for me to easily compare them between my nodes. I need to add, though, that’s really not a missing feature of their product; it is just a way to help troubleshoot my own (potentially broken) systems.

I add the tags to my own devices, not them. However, if we’ve made a mistake on our side, it’s a basic row-by-row comparison. I believe there is a way to use their SQL query feature to pull a better comparison but a method of using the GUI would be nice.

I have used this solution for about 6-8 months. For five months, I have used it as a standard user. Now, my organization created a separate admin account for me, so in total I have used it for eight months.

We have not experienced any stability issues. Other than the planned maintenance, which is short, it is always available and working great.

There have been a few very minor bugs; for instance, the auto-refresh was not working on the dashboards. When we notified them of it, they responded in less than an hour; they had replicated the issue and were working on a fix. A day later, it was done.

We have not scaled the product past the current level we are at. However, I don’t see that could ever be an issue. You just send them the flow from your devices.

If you’re scaling, you make sure your interfaces are sending the data and you're golden.

The level of technical support is beyond any vendor that I have ever worked with before.

The service is totally hosted by Kentik, with a web portal and API. I have not had issues with it being available to use. I have not tried to get to it expecting it to be available and had it not load. Occasionally we’ll get an email or pop up notification on the Web UI that planned maintenance will take Kentik down for an hour or so, these come a few days in advance of the planned service.

The only issue we have had of a technical nature was with their dashboards. Dashboards are a custom page you build and layout manually with different “Data Explorer” queries, then you turn on auto-refresh and let it continue to build the graphs as time moves on. This auto-refresh feature stopped working after an update to the Kentik UI’s look and feel. When we noticed it was not functioning we sent them an email, they responded back quickly and told us they had replicated the issue and were going to work on a fix. It was the next day when they told us to try it again, and they had indeed fixed it already! I rarely get such prompt attention to an issue.

I have used SolarWinds in another company. You get a very simple, non-configurable type of view with green, yellow, red and ingress/egress numbers. It doesn’t compare to the analytical capabilities that Kentik has.

It was set up before I joined this organization.

I am not a part of the purchasing or evaluation in any way. We still use Cacti for general stuff, but Kentik has replaced it on half of our boards so far.

While I was not a part of the implementation, if you know how to set up NetFlow on your device, just point it at Kentik. They have another setup option for a sensor that lives in your network. I have only heard of it; never used it or spoken to anyone that has.

This product is easily the best network monitor that I’ve ever seen or heard about.