Hi all! Quick question for folks running Nextflow pipelines: what are the failures you run into most often, and which ones take the longest to figure out? I'm trying to understand the most common pain points. Thanks!

Most common is running out of a resource (time or RAM), but those are typically easy to figure out and fix (slightly harder if the RAM constraint triggers weird errors in the tool, rather than a simple out-of-memory error). Harder are the things to don’t happen frequently (and are often hard to reproduce) — otherwise they would have been fixed after routine testing. One of the weirder things I have seen is a singularity container (Manta) that works on some systems inexplicably hangs (without any errors or warnings, until it times out) on others. Other (more common) hard things are when a tool gives a cryptic error, which is typically a bug in the tool, not necessarily the pipeline, and there is no obvious fix (sometimes changing parameters might help, sometimes a segfault can be avoided by filtering input data, but there is typically no deterministic way).