One to Watch: VictoriaMetrics – The story of a startup from Ukraine
This is the story of a young company seeking success, and it is a familiar story in many ways. Bright people have an idea to fill a gap in the market – check. They form a bootstrapped company – check. They work long hours hacking code – check. They spread the word and get a proof of concept out there – check. Eventually they get a sale – check. And then they build and hope to be A Big Deal – check. But VictoriaMetrics is exceptional in that it is a Ukrainian company, albeit now headquartered in San Francisco, hit hard by the outbreak of war.
Let’s talk about the company first, which has won customers including Adidas, Grammarly and the European laboratory for particle physics, CERN and which is The Stack‘s latest “One to Watch“. It makes time series database and monitoring software and and has the strapline ‘We make monitoring simple & reliable for everyone’.
What does VictoriaMetrics do?
The company’s core use case is in the disconnect between business growth and the events triggered by that growth. Metrics, for example, where VictoriaMetrics focuses, are the numerical representatives of system performance of data. Nail those and you can build a dashboard and hope to predict the future… but they move really fast and as they scale, make real demands on compute and storage.
“You can have 10% growth in your business and metrics [you track] will grow 100 or 1,000 times and that’s when companies look for alternatives,” says co-founder Roman Khavronenko.
“They start with simple solutions, hit scalability limits and still need to meet growth for next year. VictoriaMetrics was created specifically for this enormous amount of data.”
VictoriaMetrics found its first paying customer in 2019 when that customer conducted a POC, after downloading code from GitHub.
One To Watch: Paua wants to be the one card to charge them all
The company runs a fairly standard open-source model: “99% of the functionality is free to everyone and 1% is focused on enterprise, security and user management, and that is closed source and only provided to enterprise customers. The other large part of [the commercial proposition] is support and services with help and guidance for customers in running metrics. Big companies want to be sure that everything is alright. Monitoring is pretty complex and you need experts to answer your customers in how do you design a system and so on.”
VictoriaMetrics declined to comment on revenues but says it has had about 120 million downloads and its case studies include companies like website builder Wix, games platform Roblox and the writing helper app company Grammarly, which says it achieved a 10x cost reduction over its previous monitoring solution. Other users extend from Cern (for its CMS general-purpose detector system at the Large Hadron Collider) to Adidas.
VictoriaMetrics: A different way to grow
The company is hiring organically, operates lean with 10 to 20 staff and makes fairly extensive use of contractors. Unlike a Silicon Valley hotshot, it has no immediate plans to raise big funding.
“We are OK with what we get from the enterprise contracts,” Khavronenko says. “[It’s enough] to grow headcount and the company and [to develop] new solutions.”
So why not the ‘go, go, go’ of most startups?
“You promise your investors a share in your company, which is OK, but you’re losing a part of the company. Let’s imagine a case where you already have money… so why would you need additional funding? I’m not saying investment is a bad thing; it’s just a trade-off off that you’re willing to do or not willing to do.
“Not a dollar” has been raised externally although Khavronenko stresses he is not saying never to third-party funding, even if you get the impression it would have to be an out-of-the-ordinary offer.
The argument for raising money is that you can move fast, I persist, and then you gain an advantage in terms of time to market… a first-mover advantage, as they say.
One to watch: Homomorphic encryption specialist Enveil
“You can go faster but faster is not the same as quality,” he replies.
“In many cases, it’s stressful and leads to mistakes.”
For now, the company is “not willing to sacrifice [itself] to the KPIs of investors” but “it can change next week if there’s an opportunity that requires big funding. It all depends on opportunity, on cost and cost of opportunity”.
And opportunities do abound wherever there is scale of data operations. The domain is vast because “if you have huge scale you need to collect metrics from all of that”. That could be an app, your car, even your fridge: it’s all about looking at what’s happening then peeking around the corner to anticipate.
So there are opportunities to scale metrics and then there are the opportunities in adjacent market moves.
“Logs is a bit different to metrics but still one of the three pillars of observability. We’re working on this and have huge expectations this year,” Khavronenko says. The third pillar is traces and it seems natural that VictoriaMetrics would like to build a full stack in observability as its next logical move after logs. But it has lots of other stuff on its mind such as adding support for Google and Microsoft clouds after AWS and on-prem offers.
(VictoriaMetrics consists of a single small executable without external dependencies. It is available in binary releases, Docker images, Snap packages and source code. It claims to be optimized for storage with high-latency IO and low IOPS and shares disk IO graphs from benchmarks here).
Code of war
So, VictoriaMetrics knows what to do to grow and extend its reach, following in the footsteps of countless others that have followed the ancient wisdom of innovation (“build a better mousetrap and the world will beat a path to your door”) and the poetic (“a man’s reach should exceed his grasp/Or what’s a heaven for?”).
But the company has a different lens than most others, having been through a moment of war.
“It definitely impacted where we are and our lives,” says Khavronenko with characteristic understatement.
He joined the company half a year before the outbreak, leaving London, where he was working as an engineer for Cloudflare, for Kyiv: “I bought an apartment, my son was born there, then war started; that was a terrible experience,” he recalls. “Some people at VictoriaMetrics stayed, some people left. But it didn’t impact company growth. Myself and other co-workers found salvation in working more. When you can’t control the situation, you go to what you can control… and you can control what code you’re writing.”
One to watch: Cloud data warehouse startup Firebolt
Indeed. And at the time of the first crisis, there was actually a spike with more code being generated than ever.
Today, Khavronenko says he is calm and staff are safe. He says that a different sort of crisis, the Covid pandemic and lockdown, had provided a sort of trailer for what it is to work together in dispersed physical environments with restrictions in place. Today, VictoriaMetrics is a modern, dispersed startup without fixed working hours and existing on digital communications channels, GitHub for tracking tickets and the like. “We don’t have 10-to-5 [working days] and it’s totally fine if some want to sleep more, it’s what’s best for you.”
There are bigger things to think about after all and Khavronenko says he is “pretty thankful” to the EU, to countries and companies that have reached out to his countrymen and women “and helped them start their lives again”.
He currently resides in Austria. Two fellow co-founders live in the US and another in Poland.
“Everyone imagines success in their own way,” he ponders. “The great thing we’re building in VictoriaMetrics is people and a team: people who like to work with each other and build a product used by hundreds of thousands of people. Even if the product will disappear, you still have people that built it and they can build even better products and solve bigger problems. To me, that’s success: people solving real issues.”