feat(#53): Add support for resuming jobs with existing backend ID in DB

[badtst]

Resume jobs

Simple update to try and fetch status of a job on the QPU if the scheduled Job already has a backend ID

Modified

• At job creation time, the scheduler checks if job already has a valid backend id on the QPU. If so we resume execution from the status sent by the QPU
• Check job timeout in scheduler from job created_datetime returned by the QPU
  • Setting timezone information on datetime data returned by the QPU to correctly handle timeout calculation. If no tz information sent by the QPU we assume UTC.
  • By default it seems that the QPU returns UTC datetimes but no actual timezone information in the data itself

TODO

• Investigate transient error in actions: https://github.com/pasqal-io/warden/actions/runs/28670555454/job/85032615530
  • Fixed by better handling async task lifetime in test setup

[github-actions]

⚠️ critical vulnerabilities were detected in generated SBOM artifacts

SBOM vulnerability check

⚠️ Found 1 critical vulnerability entries across SBOM artifacts.

Total vulnerabilities observed: 13

warden-sbom-mariadb.cdx.json

• Severity breakdown: CRITICAL: 1, HIGH: 2, MEDIUM: 2
• Critical vulnerabilities:

Vulnerability	Severity	Component(s)	Reference
CVE-2025-65896	CRITICAL	asyncmy@0.2.11	link

warden-sbom-pg.cdx.json

• Severity breakdown: HIGH: 2, MEDIUM: 2

warden-sbom.cdx.json

• Severity breakdown: HIGH: 2, MEDIUM: 2

[MatthieuMoreau0]

By default it seems that the QPU returns UTC datetimes but no actual timezone information in the data itself

It would be useful to notify the pasqos team of this oversight. You can shoot a message on Slack #pasqos to see what they think. To clarify the risks you can mention that FC1 is hosted in Canada which is not on the same timezone as the cloud platform (europe)