Roadmap and prioritization

[mgravell]

Since Marc Gravell has now joined Redis as part of the client libraries team, we hope to see a big improvement to activity here. That means we need to get better at prioritizing. So: this is a scratch document to try and get our ducks in a row. It is an attempt to list mid-range plans. If something is missing or badly prioritized, shout!

All orders are approximate, not promises.

1. Missing commands/features

• Pending
  • hash ex
  • ACLs
  • BITFIELDs
  • XPENDING IDLE
  • client-side / server-assisted caching
  • vector search (community and enterprise)
  • active/active considerations (enterprise)
  • cluster slots
  • telemetry
• Completed recently
  • minid (merged via 2912)
  • 8.2 streaming (merged)
  • shared pub/sub (merged)
  • bitops (merged)

3. Investigate priority issues

• OOM data corruption? (example)
• possible glitch in sharded pub/sub

4. API

• refactor to use new RESP core (spiked in RESPite); remove Pipelines.Unofficial; adds new read/write API
  • also delivers new more efficient API for extending SE.Redis with custom methods (modules, etc); impacts NRedis

Current plans are:

• burn through "ready to roll" PRs to unblock features we can add quickly
• fix that OOM / corruption problem / new message API
• wider review of missing features
• profiling and optimization pass

[dceravigupta]

Are there any plan to standardize telemetry specially using Opentelemetry in SE.REDIS SDK or other client sdks maintained by Redis?

[mgravell]

added to list; needs a little thought to make sure we ship the right thing