The best way to know what it’s like to work at a company is to ask an employee. That simple idea is the driving force behind PathMotion, an online discussion platform where employee insiders share advice with job candidates. By making it easy for applicants and employees to interact, PathMotion helps to humanize the recruitment process for everyone from job candidates to human resource teams.
“What we offer is unique in the HR technology market,” says Raphaël Antonmattei, CTO at Pathmotion. “Candidates can choose to speak to employees within specific verticals and departments. And they can speak to them about anything they want. Standard recruitment content has often been reviewed so many times that it lacks personality. PathMotion provides authenticity, with honest answers from real employees.”
A microservices architecture isn’t just great for scaling, it is also an easy way to migrate to the cloud, one service at a time. For engineers who are new to the technology, this step-by-step approach is a great way to get to know the cloud and what it can do.
— Raphaël Antonmattei, CTO, Pathmotion
With clients secured in France and the UK, PathMotion is looking to expand to new markets. To do that, PathMotion collaborated with cloud competence center OP-Rate to migrate its monolithic application to a highly scalable microservices architecture on the cloud.
“A microservices architecture isn’t just great for scaling, it is also an easy way to migrate to the cloud, one service at a time,” says Raphaël. “For engineers who are new to the technology, this step-by-step approach is a great way to get to know the cloud and what it can do.”
CREATING SCALABLE, STRUCTURED ARCHITECTURE WITH MICROSERVICES
The early benefits of monolithic architecture begin to diminish as a company expands, says Raphaël: “Monolithic architecture starts off low maintenance, but as your code base grows, it gets harder and harder to maintain. It encourages an unstructured way of working; taking a lot of shortcuts in order to develop at speed. That comes at a cost, because as you add features to the code, existing features are affected. The whole thing becomes more complicated and difficult to manage.”
As PathMotion looked to develop more features and scale up its platform, it aimed to preempt any problems before they occurred. “From my experience, I knew that we would hit a block at some point in terms of our ability to maintain the system,” says Raphaël. “I wanted to anticipate it. It’s better to move when you don’t yet need to, when you still have some margin, rather than rushing things, risking mistakes, and causing stress for the company.“
PathMotion chose to build its microservices with Google Cloud, in a phased approach that would make the most of managed services. “As we first began to build our architecture, we used App Engine for some services,” says Raphaël. “It’s easy to configure, the autoscaling is in place, we don’t have to handle networking or load balancing, and it has a lot of built in security.” With its web servers still on-premises, the team began running parts of its architecture in the cloud, including its database in Cloud SQL. And to migrate the rest of its platform to the cloud, Raphaël and his team reached out to cloud competence center OP-Rate.
“OP-Rate presented us with a plan for deploying our solution on Google Kubernetes Engine (GKE),” says Raphaël. “Instead of moving everything into one single instance, it advised us to split things up and build an entire cluster. It provided two environments: a staging environment and a production environment, so there was no downtime during the migration. The project was handled very smoothly.”
The great thing about Google Kubernetes Engine is that we can allocate CPU and memory in a very granular way. That means we can put more resources on our front end, where the traffic is, and only use what we need for our back office, minimizing waste.
— Raphaël Antonmattei, CTO, Pathmotion
With microservices on GKE, all elements of the PathMotion platform are supported by dual modular redundancy, meaning that if one aspect of the solution runs into problems, it is quickly replaced. “Although we had backup servers in our previous on-premises infrastructure, they were in the same location,” says Raphaël. “Disaster recovery would take a few hours. Now we have everything replicated where we need it, so that if anything goes down, we can pick up right away, and clients don’t see a difference in the service.”
Replicating services also means Raphaël and his team can limit the impact of high-demand tasks on other processes using the same system. “We can replicate the part of the system that a lot of KPIs are pulled from, specifically for that job,” explains Raphaël. “Previously, drawing those KPIs would take hours and be very intensive on our database. Now we can call up a replica database with just two clicks, nearly eliminating the impact to other processes.” Isolating tasks and replicating resources are two of the ways PathMotion is making the most of greater granularity through GKE.
“The great thing about Google Kubernetes Engine is that we can allocate CPU and memory in a very granular way,” says Raphäel. “That means we can put more resources on our front end, where the traffic is, and only use what we need for our back office, minimizing waste.”
RESPONDING TO COMPLIANCE REQUESTS AT SPEED
The recruitment process is often sensitive, meaning privacy and security are key concerns for PathMotion and its clients. With a customer base that already includes household-name multinationals, PathMotion leverages Google Cloud to meet its compliance demands.
“Big enterprise companies present a lot of compliance and security requirements,” says Raphaël. “It’s quite hard for a small startup to fill them out at speed, and we’ve only seen the complexity and frequency of those demands increase over the years.” Because security compliance discussions typically arise late in the sales process, the PathMotion team needs to be ready to answer questions at speed, to avoid jeopardizing a deal.
“Google Cloud has a clear, well-documented security offering that meets international standards, ranging from firewalls to protection against DDoS attacks,” says Raphaël. “And when we need additional help, we can ask questions direct to the Google Cloud team.”
We want to become more structured in the way we work. The way Google Cloud is set up is ideal for this. It’s very well-structured: there’s everything from totally managed services, to services that are not managed at all. We set the cursor for where we want to be
— Raphaël Antonmattei, CTO, Pathmotion
LAYING THE FOUNDATION FOR FUTURE GROWTH
The PathMotion platform is now entirely on Google Cloud, with highly scalable microservices architecture that lays the foundation for future development. At the same time, replication on GKE makes it easy to deploy in new markets such as the US. And with its two original goals achieved, the team is now looking to use Google Cloud tools to offer AI and ML features built on top of the existing solution.
“We want to become more structured in the way we work,” says Raphaël. “The way Google Cloud is set up is ideal for this. It’s very well-structured: there’s everything from totally managed services, to services that are not managed at all. We set the cursor for where we want to be.”
