Looking into adding support for transparently compressing job input and output fields. Possibly a single algorithm to start, with more options later (e.g. zstd/brotli/gz etc.).
Should probably be similar to the way job settings are handled, i.e. where settings can be set at the queue level, or overridden per job.
Should also support separate configuration for input and output, since one/both/neither of these might benefit from compression depending on the workload type.
Looking into adding support for transparently compressing job input and output fields. Possibly a single algorithm to start, with more options later (e.g. zstd/brotli/gz etc.).
Should probably be similar to the way job settings are handled, i.e. where settings can be set at the queue level, or overridden per job.
Should also support separate configuration for input and output, since one/both/neither of these might benefit from compression depending on the workload type.