back arrow
June 1, 2021
Loretta Jones

PubMatic leverages Acceldata’s Data Observability platform to optimize performance and cost at massive scale

PubMatic leverages Acceldata’s Data Observability platform to optimize performance and cost at massive scale

PubMatic is one of the United States largest AdTech companies. Since 2006, PubMatic has created an efficient global infrastructure with eight global data centers. The company is one of the industry’s leaders in  programmatic advertising innovation.

As of Dec 2020, PubMatic  served 171 billion ad impressions, handled one-trillion advertiser bids, and processed more than 2PB of new data on a  daily basis.


PubMatic is in hyper-scale mode. Their current environment includes 3000+ nodes, 150+ Petabyes and 65+HDPs (Horton Dataworks Platform) Clusters and is expanding rapidly. In addition, PubMatic usesYarn, Kafka (50+ small Kafka clusters with 10-15+ nodes/cluster), Spark, HBase and open source - HDP (Hortonworks Data Platform).


Because of their massively scaled environment, PubMatic consistently experienced  high MTTR (Mean Time to Resolution) metrics, frequent outages, and performance bottlenecks.

Many of the issues stemmed from large numbers nodes. The system’s instability resulted in time-consuming operational issues and constant daily firefighting. In addition, PubMatic was looking for ways to reduce its  infrastructure and OEM support costs.

Business Impact

When PubMatic’s data system performance wasn’t able to keep pace with its rapidly expanding business requirements,  the company decided to implement  a data observability platform to improve data operations reliability, scalability, and return on investment.

The inability to correlate events across the infrastructure, data layers and pipelines meant that PubMatic could not materially improve  their ‘cost per ad impression’ metric, which is one of their most critical performance metrics.

In addition, the company’s rapid scaling resulted in unnecessary software licenses, which they felt they could better align with actual needs. Finally, engineering’s constant involvement in resolving operational system issues caused a distraction from the real objectives of scaling the data system to support the fast-growing business requirements.


PubMatic began using Acceldata’s Pulse product in mid-2020. At the data compute layer, Pulse immediately provided  improved visibility into the inner-workings of PubMatic’s data applications and comprehensive observability for complex, interconnected data systems.

One of Pulse’s most important benefits was its ability to  predict, prevent and optimize PubMatic’s data system performance at the very large scale that today’s digital ad market requires.

In PubMatic’s environment, Acceldata Pulse isolated bottlenecks and automated performance improvements. The product distinguished between mandatory and unnecessary data to ensure scaled growth that could reliably support all critical enterprise and customer-facing analytics requirements. Acceldata Pulse has helped PubMatic:

  • Reduce ‘cost per ad impression’ - a key performance metric
  • Improve reliability of data pipelines
  • Eliminate of day-to-day engineering involvement and firefighting on outages and performance degradation issues
  • Decrease  OEM support costs
  • Optimize HDFS to reduce block footprint by 30%
  • Consolidate Kafka cluster and saved infrastructure costs
  • Saved millions of dollars in unnecessary software licenses

“Acceldata provided the data observability tools and expertise to make our data pipelines more reliable. They helped us optimize HDFS performance, consolidate Kafka clusters, and reduce cost per ad impression, which is one of our most critical performance metrics. Acceldata's data observability saved us millions of dollars for software licenses that we no longer need. Now we can focus on scaling to meet the needs of rapidly growing business.”

Ashwin Prakash, Engineering Leader, PubMatic