Picking Pearl from Seabed: Extracting Artefacts from Noisy Issue Triaging Collaborative Conversations for Hybrid Cloud Services

Site Reliability Engineers (SREs) play a key role in identifying the cause of an issue and preforming remediation steps to resolve it. After is reported, SREs come together virtual room (collaboration platform) triage issue. While doing so, they leave behind wealth information, form conversations, which can be used later for triaging similar issues. However, usability these conversations offer challenges due them being scarcity conversation utterance label. This paper presents novel approach artefact extraction from noisy with minimal labelled data. We propose combination unsupervised supervised models human intervention that leverages domain knowledge predict artefacts small amount data use fine-tuning already pre-trained language model prediction on large Experimental results our dataset show proposed ensemble better than using either one individually. also present deployment case study prediction.