Synapse Dedicated SQL Pool
Does ZopNight manage Synapse Dedicated SQL Pool?
Synapse dedicated SQL pools bill by provisioned DWU block continuously while online, commonly thousands of dollars a month, and pausing halts DWU compute billing entirely. ZopNight pauses pools between load windows on schedule, guided by 60 days of utilization metrics. Storage keeps billing while paused, and a pause terminates running queries.
Rules that fire on Synapse Dedicated SQL Pool
No active rule family targets Synapse Dedicated SQL Pool today. Rules that used to are retired, and retired rules publish no pages and fire no findings. Scheduling and permissions coverage are unaffected.
At a glance
| Field | Value |
|---|---|
| Scheduling notes | paused and resumed via the Synapse pool Pause/Resume operations, halting DWU compute billing while paused. |
Dedicated SQL pools provision data-warehouse compute in DWU blocks that bill continuously while online, often thousands of dollars monthly. Pausing outside load windows is the single biggest Synapse saving.
DWU blocks bill until someone pauses them
A dedicated pool’s meter is its provisioned DWU level, charged for every hour the pool is online regardless of whether a single query runs. Warehouse workloads are naturally bursty: a nightly load, a morning refresh, ad-hoc analysis in office hours. That makes the always-online default an expensive mismatch. Storage is metered separately and survives a pause, so pausing converts a large continuous compute charge into a modest storage-only floor.
Pause and Resume as scheduled operations
ZopNight uses the Synapse pool Pause and Resume operations, halting DWU compute billing while paused. The blunt edge is that pausing terminates running queries: a pause landing mid-load kills the load. Schedules should therefore bracket the known load windows with margin, and resume early enough that the pool is warm before the first job of the day. Between load windows is exactly where schedules pause pools, the pattern that captures most of the available saving.
Signals ZopNight reads on each pool
Discovered via Azure Resource Graph with DWU sizing and state. Azure Monitor metrics (60-day lookback) show utilization, Cost Management billing attributes spend, and schedules pause pools between load windows. ML auto-tagging covers this type. The DWU-level and state data together answer the two review questions that matter: how big is this pool, and how much of the day is it actually working?
The usual dedicated-pool leaks
Pools resumed for a one-off backfill and never re-paused. DWU levels set for quarter-end reporting peaks and left there year-round. And duplicate environments (dev, test, UAT) each carrying its own always-online pool when a single scheduled one would serve.
Pool state in the Synapse portal blade
Azure portal → Azure Synapse Analytics → select the workspace → SQL pools: each pool row shows its status, Online or Paused, next to its DWU size. Online is the state that bills; the Pause button, or a ZopNight schedule driving the same operation, is what ends it.