auth token expired mid-data quality run error failed checkpoint

TL;DR

A credential that expires mid-run must be refreshed, not mourned: give the agent a refresh path it can use without a human, such as a rotating credential from its vault or a workload identity, check the credential's remaining life before starting a long run, and make the run resumable so an expiry costs a pause, not the whole checkpoint. Runs longer than the credential's life will always hit this until the refresh exists.

auth token expired mid-data quality run error failed checkpoint

Use this when

• The check or run fails with this symptom: the run fails partway with an authentication error after the credential expires.
• You are working on an agent data quality run failing when its credential expires mid-checkpoint and the failure repeats on the same data or configuration.
• You need the failing rows, rule, or stage isolated before changing thresholds or code.

Not for this skill when

• An authentication failure at run start, which is a configuration problem rather than a mid-run expiry instead.
• The error text names a different tool, stage, or root cause. Match the verbatim error first and use that skill.

Steps

1. Determine the credential type and its lifetime, and compare it with the run's typical duration. Verify: the lifetime-versus-duration gap is quantified
2. Check how the agent obtains the credential at start and whether any refresh mechanism exists in that path. Verify: the acquisition and refresh paths are mapped
3. Implement refresh through the vault or identity provider the deployment already uses, triggered before expiry. Verify: refresh happens automatically ahead of expiry
4. Add a pre-run check that refuses to start a run longer than the credential's remaining life without a working refresh. Verify: the pre-run check is in place
5. Run a checkpoint longer than the original lifetime and confirm it completes across a refresh. Verify: the long run completes across the refresh boundary

Variant phrasings

auth credential expired mid data quality run checkpoint

The verbatim mid-run expiry phrasing, with the credential named generically.

agent run longer than credential lifetime refresh

The refresh-path phrasing for vault or workload identity.

data quality run authentication failed partway expiry

The partway-failure phrasing for expiry during execution.

Compatibility: Agent data quality runs using warehouse or Great Expectations credentials with limited lifetimes

Why it happens

Short-lived credentials are deliberate security policy, and long validation runs are deliberate thoroughness; the two collide by design. Agents that fetch a credential once at start inherit a hard deadline that has nothing to do with the data, and only a refresh path resolves the collision.

Edge cases / pitfalls

• Never extend credential lifetimes to fit the run when a refresh path is available; lifetime is a security control.
• Make validation runs resumable regardless, so any interruption, not only expiry, is cheap.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_A4KoX38I4Dx7v1p8tZRzQw