For staters, consider what contributes to the cost in this case:
- Compute (vCore) charges — stopped instances do not incur compute (vCore) charges while in the stopped state. This is the primary source of savings from a nightly stop/start schedule.
- Storage charges (provisioned data storage) — continue to be billed while stopped.
- Backup storage & backup/retention charges — continue to be billed (local or long-term retention).
- Other fixed charges (e.g., networking egress, IP addresses) — generally continue if applicable.
- Licensing — included in the pay-as-you-go vCore rate, so pausing compute pauses that part of the charge.
Inputs you need (get from Azure Pricing Calculator / portal):
R = compute price per vCore-hour in Norway East (pay-as-you-go) — e.g., in USD/hour (or your billing currency)
v = number of vCores (you said 4)
S_GB = provisioned storage (GB) — you said 64 GB
P_storage = storage price per GB-month (same region)
P_backup = backup storage price per GB-month (if applicable)
h_stop_per_day = hours stopped per day (you said 11:00 pm → 06:00 am = 7 hours)
days_month = days in month (use 30 or exact month days)
hours_month = 24 * days_month (typical use: 730 or 720; choose one)
Formulas:
Hours stopped per month: H_stop = h_stop_per_day * days_month
Compute savings per month: Savings_compute = R * v * H_stop
Compute what fraction of compute is saved: Pct_compute_saved = H_stop / hours_month (e.g., 210/720 = 29.17%)
Net monthly cost change = (compute bill without schedule) − (compute bill with schedule) But since storage & backup continue, they aren’t changed:
Compute_monthly_no_schedule = R * v * hours_month
Compute_monthly_with_schedule = R * v * (hours_month - H_stop)
Storage_monthly = P_storage * S_GB
Backup_monthly ≈ P_backup * backup_GB (use your retained backup GB)
- Total monthly savings =
Savings_compute(since only compute changes).
To get the accurate regional prices, you can use the following options:
- Azure Pricing Calculator — select Azure SQL Managed Instance (or your product), choose Norway East, Pay-as-you-go, and read the vCore/hour and storage per GB/month lines.
- Azure Portal / Cost Management — view current unit prices or your subscription’s meter rates (use actual invoices or the Cost Analysis blade to see meter rates).
- Azure RateCard / Pricing APIs — if you automate, use the Azure Retail Prices API or RateCard to programmatically pull
unitPricefor your region/meter.
There are some additional caveats to consider:
- Minimum/rounding billing increments: verify whether compute is billed in per-minute or per-second increments in that SKU (usually hourly-based rates are reported but charge may be prorated by the second). Use your invoice or cost API to confirm exact granularity.
- Start/stop automation timing — ensure stop occurs after all nightly jobs complete; a start during heavy load may cause performance/availability impacts.
- SLA/HA — frequent stop/start is fine for dev/test but may not be appropriate for production replicas or HA setups.
- Backup retention — long-term retention could add substantial monthly backup storage charges that are unaffected by stopping compute.
- Cold start latency — starting the instance reinitializes resources and may take time; plan windows accordingly.
If the above response helps answer your question, remember to "Accept Answer" so that others in the community facing similar issues can easily find the solution. Your contribution is highly appreciated.
hth
Marcin