Monitoring bacterial communities in a drinking water treatment plant (DWTP) may help to understand their regular operations. Bacterial community dynamics in an advanced full-scale DWTP were analyzed by 16S rRNA metabarcoding, and microbial water quality indicators were determined at nine different stages of potabilization: river water and groundwater intake, decantation, sand filtration, ozonization, carbon filtration, reverse osmosis, mixing chamber and post-chlorination drinking water. The microbial content of large water volumes (up to 1100 L) was concentrated by hollow fiber ultrafiltration. Around 10 million reads were obtained and grouped into 10,039 amplicon sequence variants. Metabarcoding analysis showed high bacterial diversity at all treatment stages and above all in groundwater intake, followed by carbon filtration and mixing chamber samples. Shifts in bacterial communities occurred downstream of ozonization, carbon filtration, and, more drastically, chlorination. Proteobacteria and Bacteroidota predominated in river water and throughout the process, but in the final drinking water, the strong selective pressure of chlorination reduced diversity and was clearly dominated by Cyanobacteria. Significant seasonal variation in species distribution was observed in decantation and carbon filtration samples. Some amplicon sequence variants related to potentially pathogenic genera were found in the DWTP. However, they were either not detected in the final water or in very low abundance (<2%), and all EU Directive quality standards were fully met. A combination of culture and high-throughput sequencing techniques may help DWTP managers to detect shifts in microbiome, allowing for a more in-depth assessment of operational performance.