Our client is an ads distribution agency. The receive millions of clicks on daily basis on their ads. Click recording stats are very important for the business and it needs to be presented to advertisers and calculation of RoI. On advertiser dashboard different aspects of clicks were presented in charts. As all this data was stored in MySQL so the queries which were being used for charts were slow.
We designed and implemented Elasticsearch to overcome the issues being faced due to amount of data. We created shards based on the date and used Elasticseach aliases to provide monthly and weekly views. Indexing all the existing data into Elasticsearch was another challenge. We used our product Griffin to index all existing data into Elasticsearch at fast pace.
Our this solution can resolve issues where inserting frequent data into traditional database is a bottleneck. Also solves issues where search needs to be provided on large volume of data.