PicsArt’s Technology Landscape
PicsArt enables its creative user base to freely upload or draw, edit, and exchange digital images in a worldwide, web-based community. PicsArt works on the web and on Android, iOS, and Windows Phone platforms.
According to Mikayel Vardanyan, Chief Technology Officer, PicsArt, PicsArt apps and technologies process millions of user requests each day. “We will also soon provide a public API to all of our developers, giving them an opportunity to use our backend services to create their own apps,” says Vardanyan. Though this will satisfy customer requests for additional functionality, it will also further multiply the intricacies of the PicsArt systems, making it ever more critical that they perform properly from foundation to pinnacle.
PicsArt Experiences the Same Pitfalls as Other Tech Firms, and Responds
As with any technological system, PicsArt’s infrastructure is at risk of downtime, which looms ready to strike the business and to challenge customer satisfaction at any time. Unforeseeable events as well as planned maintenance outages create downtime and are an expected part of the systems lifecycle, even though they are not always welcome.
But malfunctions and downtime can exacerbate DevOps processes and corrupt and diminish the user experience. Downtime can temporarily take out entire services, ones that customers will definitely notice and miss when they become unavailable. “If, for example, the users cannot upload their pictures, that would certainly be bad for us,” says Vardanyan.
Sufficient downtime on the PicsArt backend systems could also cause user registrations functions to fail. The backend also supports PicsArt’s community activities such as sharing artwork, liking others’ artwork, and commenting on people’s creations, which could suffer. “We try to prevent downtime by providing redundancy and scalable systems. We properly test our systems on a regular basis,” says Vardanyan.
“When we start to sense issues such as these, our operations team goes to work immediately to fix what we need to fix. We investigate all the issues we come across and respond by adding more layers of redundancy and scalability,” says Vardanyan.
PicsArt Sought Picture Perfect Improvements from APM
PicsArt was looking for increasing visibility into its infrastructure in order to monitor for the aforementioned problems at their very outset, to create alerts on trouble issues and status reports, and to derive metrics about what is happening across the entire infrastructure from nation to nation, and region to region.
PicsArt knew that it needed extensive monitoring and that the best way to go would be to have that monitoring centralized in a single product. “We needed to monitor the infrastructure 24/7 in order to provide the best possible service for our community,” says Vardanyan; “we needed external as well as internal monitoring and we needed to monitor the user experience”.
APM Frames Performance Challenges Nicely
PicsArt currently uses Monitis. At PicsArt, systems are continually wrestling with pressing loads and with load balancing. If systems performance should lag for any reason, the network can take center stage as the weakest link and the strongest bottleneck. “We utilize the different alerting capabilities of Monitis, such as voice calls and SMS, using all the escalation procedures that its notification systems offer, too. We also use Monitis to create custom monitors for certain scripts and for Redis jobs,” says Vardanyan.
The cloud-based monitoring solution enables the digital art editing and sharing platform provider to leverage views from a number of global vantage points, clarifying transparency into PicsArt infrastructure. “With Monitis, we can check our global CDN locations and find out what location may have an issue. The benefit is that a potentially issue-stricken location does not escape the monitoring routine,” says Vardanyan. The sooner the enterprise sees it, the faster it can fix it.
The conveniences and efficiencies of dissecting it all inside a single dashboard enable PicsArt to repair the real problems in its DevOps and platform delivery processes without touching anything that is not out of tune.