[BICJONSQLPOOL / DW300c] Can the pool safely run two additional daily pipelines (noon & 6 PM)? — DWU used reads 300 with no active requests

Subject: BICJONSQLPOOL (DW300c): capacity check before adding noon + 6 PM pipeline runs, and DWU-used-at-300 with no active requests

Hello,

We run a Dedicated SQL Pool named BICJONSQLPOOL at service level DW300c. We are planning to add a pipeline that would run twice daily - around 12:00 PM (noon) and 6:00 PM - and want to confirm whether the pool can absorb this without contention, before we schedule it.

We have two open questions and would appreciate a backend-level assessment.

=== Question 1: Why is DWU used pinned near 300 with no active requests? ===

• The "DWU used" metric sits at or near the 300 limit for most of the afternoon, and daily aggregated statistics show it near 300 almost constantly.
• However, sys.dm_pdw_exec_requests shows no active user requests running during those times (no Running or Suspended queries; a manual check returned only our own monitoring query with ~17 ms queue wait).
• The "Active requests" metric is near zero for almost the entire period, with a single ~220 spike at ~8:41 AM that does NOT correlate with the sustained high DWU used readings.

We understand DWU used is documented as only a high-level approximation, but we need to know whether BICJONSQLPOOL is genuinely resource-constrained in the afternoon or whether this is a reporting artifact, since it directly affects whether we can add afternoon/evening load.

=== Question 2: Current pipeline schedule and proposed additions ===

All of our existing pipeline runs currently execute in the OVERNIGHT / EARLY-MORNING window (approx. 12:47 AM to 8:06 AM). There is currently NO pipeline activity in the afternoon or evening. Representative runs from 8/6/2026 are listed below.

We want to know if adding two runs at ~12:00 PM and ~6:00 PM is safe.

Existing runs (8/6/2026):

| Pipeline | Start | Duration | Result |

|---|---|---|---|

| ITSM_AUTODB_PRD_Pipeline | 8:00:01 AM | 6m 7s | Success |

| AI Code Review_Pipeline_DEV | 7:45:00 AM | 3m 10s | Success |

| IPOADM_Pipeline_PRD | 7:30:00 AM | 2m 33s | FAILED |

| CUSTOMFIELDVALUE_Pipeline_DEV | 7:15:00 AM | 5m 56s | Success |

| ORACLE_MSG_BIZWBP_PRD | 6:46:00 AM | 2m 16s | FAILED |

| JiraSQL_Insight_Pipeline_DEV | 6:45:00 AM | 9m 38s | Success |

| ITSM_DEVOPS | 6:30:01 AM | 43m 54s | FAILED |

| JiraSQL_Pipeline_DEV | 6:15:00 AM | 11m 14s | Success |

| OY_WikiSQL_Pipeline_PRD | 6:05:00 AM | 3m 10s | Success |

| OY_JiraSQL_Pipeline_PRD | 6:05:00 AM | 2m 30s | Success |

| CUSTOMFIELDVALUE_Pipeline_PRD | 5:15:01 AM | 2h 38m 38s | Success |

| JiraSQL_Pipeline_PRD | 5:15:01 AM | 12m 48s | Success |

| WikiSQL_Pipeline_PRD | 5:10:00 AM | 4m 3s | Success |

| ICMADM_Pipeline_PRD | 5:01:01 AM | 2m 14s | FAILED |

| SQLMI_JSP_PRD_PIPELINE_PRC | 4:05:01 AM | 10m 30s | Success |

| REMEDY_Pipeline_PRD | 4:00:00 AM | 11m 4s | Success |

| JiraSQL_Insight_Pipeline_PRD | 12:47:06 AM | 15m 43s | Success |

Proposed additions:

• New pipeline run #1: ~12:00 PM daily
• New pipeline run #2: ~6:00 PM daily

Specific questions:

A. Given that DWU used already reads ~300 in the afternoon despite no active requests, would scheduling real workload at 12:00 PM and 6:00 PM actually encounter contention, or is the 300 reading not reflective of available capacity?

B. Is there enough headroom on DW300c to run these two additional pipelines in the afternoon/evening, given the existing overnight batch load and the long-running CUSTOMFIELDVALUE_Pipeline_PRD (2h 38m)?

C. Do you recommend any workload management configuration (resource classes / workload isolation) before adding afternoon load, rather than scaling the pool up?

Pool details:

• Pool name: BICJONSQLPOOL
• Service level: DW300c
• (Please let us know any workspace/subscription details you need.)

Thank you,

[Your name]