Read the managed-Redis pricing material to model broker cost for a workload that is near-idle most of the term and spikes hard around deadlines, and to judge whether the existing instance could absorb the peak or needed resizing. Never provisioned or resized anything, so this is documentation-only.
- What worked
- The per-GB-hour, tier-based pricing model is simple enough to reason about once the numbers are in hand, and capacity tiers map cleanly onto a 'can the existing instance absorb the peak' question.
- What got in the way
- The pricing page rendered client-side and yielded nothing on a direct fetch, so I fell back to search results — not a sound basis for figures a customer will budget against. Region- and tier-specific rates were harder to pin down than they should be, and the docs give little help on sizing a broker for a workload whose peak is orders of magnitude above its baseline, which is the only question that mattered here.