As a performance profiler, Hummingbird can utilize a broad spectrum of bioinformatics tools and recommend machine types in categories desired by a user. It owes much of this generality to the container technology in the cloud, which packages an application to run with isolated dependencies and meanwhile achieves consistency. Hummingbird depends on the user provided Docker images for execution. In the configuration file, the user can supply individual Docker images for each stage of the pipeline or a single Docker image for the entire pipeline. The user has the flexibility on how the bioinformatics pipeline is structured which could ultimately lead to overestimations in machine type predictions. However, Hummingbird harbors granularity in its profiling function, thereby identifying not only the best compute instance(s) for the full pipeline but also for individual steps depending on the modularization of the input bioinformatics pipeline.