Idle Azure ML Compute
A compute instance bills for its VM size until it is stopped, and it does not stop itself without an idle-shutdown policy.
Free to start. No card. The playground just needs your work email.
Azure
Found, explained, fixed.
Findings with a dollar figure attached, idle, oversized, orphaned, unscheduled and undiscounted spend.
Detect
ZopNight checks this automatically across Azure, with read-only access to the account.
Explain
Every finding says exactly what to change: Set minimum nodes to zero on the ML compute cluster. Where the saving can be proven it is priced; where it cannot, the finding says so.
Fix
After a guided review, where you confirm the change, ZopNight pauses the service, then checks the change took.
- Applies to
- ML Compute Clusters on Azure
- The fix, by hand
- Open Azure Machine Learning studio at ml.azure.com. Note this is the current product; ML Studio (classic) sunset 2024-08-31 and shouldn't be in use.
- Compute → Compute clusters → select the cluster behind the resource.
- Review Jobs over the last 14 days. Confirm no scheduled training pipelines.
- Edit cluster → set Minimum number of nodes = 0 (this is the lever; the cluster scales to zero when idle and back up on demand).
- Save; verify within 30 minutes that the cluster has scaled down.
- If the cluster is genuinely abandoned, delete it instead.
ZopNight does this for you after a guided review.
- Category
- Idle resources. Resources running but doing no useful work: zero invocations, no connections, no traffic.
- Where it appears
- The Savings tab of Recommendations, with every affected resource listed.
- Rule ID
RC-223- Full reference
Related checks
See the idle resources in your account.
Connect a read-only role and the first pass runs on your own estate. This check, and the rest of the catalogue, with it.
Prefer to talk it through first? Book 20 minutes with the team.
- $30M+annualised cloud spend under management
- 550K+resources tracked since launch
- 20-60%off the bill in the first month
- SOC 2Type II report, plus ISO 27001
Figures published on zop.dev.