Distributed systems have become ubiquitous in recent years, with those based on distributed ledger technology (DLT), such as blockchains, gaining more and more weight. Indeed, DLT ensures strong data integrity thanks to complex cryptographic protocols and high distribution. That said, even the most powerful systems will never be perfect, and, in fact, the larger they get, the more exposed they become to threats. For traditional systems, log auditing effectively addresses the problem and makes it possible to analyze the use of applications. However, DLT systems still lack a wide range of log analyzers due to the particularities of their distribution. To help remedy this weakness, we propose here a generic auditing system called DELTA (for Distributed Elastic Log Text Analyzer). By coupling Natural Language Processing with the Docker Engine of the Filebeat, Logstash stack, Elasticsearch and the visual tool Kibana, DELTA tracks, analyzes and classifies logs generated by DLT systems. Additionally, it enables real-time monitoring thanks to visual analysis and querying of structured data. DELTA is the first auditing system applicable to blockchains that can be integrated with the Docker Engine. In addition to describing its general principles and specific components, we illustrate its application to Hyperledger Fabric, the most popular of the platforms for building private blockchains.