@Sammy Human fasta + indexes will be very large and take up space on disk, whether you create these indexes yourself (Indexing reference genomes with Data Managers: Resources, tutorials, troubleshooting) or import pre-compiled indexes (http://datacache.galaxyproject.org/).
If you do have space for the indexes, there may be downstream problems if there is limited RAM. Meaning, tools can fail later for memory resources. Tools use the same amount of RAM whether run in Galaxy or line-command. How much RAM a tool uses depends on a few factors: how the tool itself handles data, tool parameters, input sizes, input content…
See the underlying tool’s manual for how much RAM is recommended – most will have a link to the source tool’s home page or try a google search.