Read the visible stage and reason
A failed crawl, analysis, provider request, or import should expose its stage and message. Capture the project, time, action, and message before retrying so repeated failures can be compared.
Preserve completed evidence
A later analysis can fail after the crawl snapshot is saved. Retry the failed layer instead of deleting a good crawl. Similarly, a connection refresh should not erase earlier imported records simply because one provider page was incomplete.
Avoid duplicate paid work
Check whether the previous provider run completed before pressing Run again. Locked tools and failed requests should state whether a provider call occurred and whether a saved result was created.
Escalate with useful diagnostics
Send the exact page, workspace site, local time with timezone, action, visible message, and whether a retry changed the state. Do not send passwords, API keys, access tokens, or customer private data.