Is your feature request related to a problem? Please describe.
neodb's use of redis queues for workload scheduling unnecessarily requires extra Docker containers with dedicated RAM, causing frequent out-of-memory errors on 4GB nodes that seem unwarranted by the relatively modest workload. As a workload mechanism, redis fails to smooth out resource consumption but puts additional stress on RAM instead and also creates unnecessary friction with reconnects and Docker dependencies.
Describe the solution you'd like
Use a Python wrapper lib for redis MQ that would, in a second step, allow switching redis for Postgres message queues, eliminating the need for running redis or other unnecessary in-memory data storage (except maybe Django core caching).
Is your feature request related to a problem? Please describe.
neodb's use of redis queues for workload scheduling unnecessarily requires extra Docker containers with dedicated RAM, causing frequent out-of-memory errors on 4GB nodes that seem unwarranted by the relatively modest workload. As a workload mechanism, redis fails to smooth out resource consumption but puts additional stress on RAM instead and also creates unnecessary friction with reconnects and Docker dependencies.
Describe the solution you'd like
Use a Python wrapper lib for redis MQ that would, in a second step, allow switching redis for Postgres message queues, eliminating the need for running redis or other unnecessary in-memory data storage (except maybe Django core caching).