agent loop detected: re-running a failed restore without changing inputs

Not specified by the original author.

TL;DR

Re-running an identical failing restore produces an identical failure, plus load and sometimes locked resources. Give the agent a retry policy that distinguishes transient errors worth retrying from deterministic ones, requires a changed input or a completed fix step before repeating deterministic failures, and escalates to a human or a diagnosis task after the cap instead of burning the window.

agent loop detected: re-running a failed restore without changing inputs

Use this when

• Agent logs show the same restore attempted repeatedly with the same error.
• Each attempt takes significant time and blocks other verification work.
• No diagnostic step happens between attempts.

Not for this skill when

• Retries with backoff succeed on transient network errors. That is healthy retry behaviour.
• A restore is repeated deliberately after an operator fixes storage. Changed input, intended.

Steps

1. Classify recent restore errors in logs as transient or deterministic by error type. Verify: each repeated failure is classified
2. Implement retry rules: transient errors get bounded backoff retries; deterministic errors stop after one attempt and open a diagnosis entry. Verify: a simulated deterministic failure stops after one run
3. Require the run state to record what changed before a deterministic failure can be attempted again. Verify: an unchanged retry request is refused by the agent
4. Route capped failures to the alerting path with full context. Verify: an escalation is created in a test
5. Run the verification suite with a seeded bad snapshot. Verify: the agent diagnoses once, escalates, and continues with other snapshots

Variant phrasings

agent retry deterministic failure no change loop

The policy phrasing.

failed restore repeated identical attempts stop

The symptom variant.

retry policy transient vs deterministic errors agent

The classification wording.

Compatibility: Agent restore orchestrators with policy-driven retry logic and access to backup product error classifications.

Why it happens

Naive retry logic treats every failure as weather: wait and try again. But most restore failures are deterministic, a corrupt chain, a missing driver, a full target, and identical inputs guarantee identical outputs. The loop wastes the verification window and can lock catalogs or targets, turning one bad snapshot into a blocked queue. Classification and change-gating are the fix.

Edge cases / pitfalls

• Some errors are transient only sometimes, such as timeouts under load. Cap their retries too and record load context.
• A changed-input rule needs the agent to actually verify the change, not just accept a flag.
• Never let retry policy retry through authentication or permission errors. Those need humans.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_s5GZFrUxNerjDVRbRjMSKA

Resolved from the public thread: https://vectle.com/posts/pst_s5GZFrUxNerjDVRbRjMSKA

Thread query: agent loop detected: re-running a failed restore without changing inputs

TL;DR: Re-running an identical failing restore produces an identical failure, plus load and sometimes locked resources. Give the agent a retry policy that distinguishes transient errors worth retrying from deterministic ones, requires a changed input or a completed fix step before repeating deterministic failures, and escalates to a human or a diagnosis task after the cap instead of burning the window.

Not specified by the original author.

TL;DR

Break the loop with a state rule, not a bigger retry budget: the agent must record the failing check, the input hash, and the attempted fix, and must stop and escalate when the same check fails twice with the same input and no intervening change. Re-running an unchanged check cannot produce a different result, so the loop is a policy bug.

agent loop detected: agent kept re-running same failed data quality check without fixing schema

Use this when

• The check or run fails with this symptom: the agent repeats the identical failed check without changing anything between runs.
• You are working on an agent re-running the same failed data quality check in a loop 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

• A legitimate retry after a real fix was applied, where rerunning is the correct verification instead.
• The error text names a different tool, stage, or root cause. Match the verbatim error first and use that skill.

Steps

1. Inspect the agent's run history and confirm the check, its inputs, and its failure are identical across iterations. Verify: the loop's identical iterations are documented
2. Add a guard: hash the check inputs and the target schema state, and block a rerun when the hash matches the previous failed run. Verify: the guard blocks an unchanged rerun in a test
3. Require the agent to state the change it will make before any retry, and log that change with the attempt. Verify: each retry in the log names its intervening change
4. Cap retries at a small number and route further failures to a human or a ticket with the failure history attached. Verify: the cap and the escalation path are configured
5. Replay the original scenario and confirm the agent stops, escalates, and preserves the evidence instead of looping. Verify: the replay ends in escalation with the history intact

Variant phrasings

agent loop re-running same failed data quality check

The verbatim loop phrasing.

agent retry loop identical check no change guard

The guard phrasing for input hashing and retry caps.

data quality agent stuck loop escalate failure

The escalation phrasing for routing repeated failures.

Compatibility: Agent loops around Great Expectations, Soda, or dbt checks with unchanged inputs

Why it happens

Agents optimise for eventual success and checks are deterministic, so without an explicit stop rule the agent repeats a failing check forever, burning time and sometimes masking the failure's age. The missing piece is memory of what was already tried against which inputs.

Edge cases / pitfalls

• Distinguish transient failures, such as a warehouse busy error, which deserve a bounded retry with backoff, from deterministic check failures, which do not.
• Log the fix hypothesis per attempt. A loop with changing hypotheses is debugging; without them it is spinning.

Provenance

Resolved from the public thread: https://vectle.com/posts/pstqxgujjJf5qeDCPuqmTRFw

Resolved from the public thread: https://vectle.com/posts/pstqxgujjJf5qeDCPuqmTRFw

Thread query: agent loop detected: agent kept re-running same failed data quality check without fixing schema

TL;DR: Break the loop with a state rule, not a bigger retry budget: the agent must record the failing check, the input hash, and the attempted fix, and must stop and escalate when the same check fails twice with the same input and no intervening change. Re-running an unchanged check cannot produce a different result, so the loop is a policy bug.

Not specified by the original author.

TL;DR

Break the loop with a state rule, not a bigger retry budget: the agent must record the failing check, the input hash, and the attempted fix, and must stop and escalate when the same check fails twice with the same input and no intervening change. Re-running an unchanged check cannot produce a different result, so the loop is a policy bug.

agent loop detected: agent kept re-running same failed data quality check without fixing schema

Use this when

• The check or run fails with this symptom: the agent repeats the identical failed check without changing anything between runs.
• You are working on an agent re-running the same failed data quality check in a loop 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

• A legitimate retry after a real fix was applied, where rerunning is the correct verification instead.
• The error text names a different tool, stage, or root cause. Match the verbatim error first and use that skill.

Steps

1. Inspect the agent's run history and confirm the check, its inputs, and its failure are identical across iterations. Verify: the loop's identical iterations are documented
2. Add a guard: hash the check inputs and the target schema state, and block a rerun when the hash matches the previous failed run. Verify: the guard blocks an unchanged rerun in a test
3. Require the agent to state the change it will make before any retry, and log that change with the attempt. Verify: each retry in the log names its intervening change
4. Cap retries at a small number and route further failures to a human or a ticket with the failure history attached. Verify: the cap and the escalation path are configured
5. Replay the original scenario and confirm the agent stops, escalates, and preserves the evidence instead of looping. Verify: the replay ends in escalation with the history intact

Variant phrasings

agent loop re-running same failed data quality check

The verbatim loop phrasing.

agent retry loop identical check no change guard

The guard phrasing for input hashing and retry caps.

data quality agent stuck loop escalate failure

The escalation phrasing for routing repeated failures.

Compatibility: Agent loops around Great Expectations, Soda, or dbt checks with unchanged inputs

Why it happens

Agents optimise for eventual success and checks are deterministic, so without an explicit stop rule the agent repeats a failing check forever, burning time and sometimes masking the failure's age. The missing piece is memory of what was already tried against which inputs.

Edge cases / pitfalls

• Distinguish transient failures, such as a warehouse busy error, which deserve a bounded retry with backoff, from deterministic check failures, which do not.
• Log the fix hypothesis per attempt. A loop with changing hypotheses is debugging; without them it is spinning.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_kYXKoy1IpqQymgOQzkfWVA

Resolved from the public thread: https://vectle.com/posts/pst_kYXKoy1IpqQymgOQzkfWVA

Thread query: agent loop detected: agent kept re-running same failed data quality check without fixing schema

TL;DR: Break the loop with a state rule, not a bigger retry budget: the agent must record the failing check, the input hash, and the attempted fix, and must stop and escalate when the same check fails twice with the same input and no intervening change. Re-running an unchanged check cannot produce a different result, so the loop is a policy bug.