Interactive analysis notebooks on DESY batch resources

Abstract Batch scheduling systems are usually designed to maximise fair resource utilisation and efficiency, but less well for demanding interactive processing, which requires fast access resources while low upstart latency is only of secondary significance high throughput performance computing systems. The clusters at DESY intended as batch end users run massive analysis simulation jobs enabling turnaround systems, in particular when processing expected feed back operation instruments near real-time. continuously increasing popularity Jupyter Notebooks online made an integration this technology into the indispensable. We present here our approach utilise HTCondor SLURM backends integrate Notebook servers techniques involved provide access. chosen offers a smooth user experience allowing customize allocation tailored their computational requirements. In addition, we outline differences between HPC HTC implementations give overview running services.