ProviderStatus workflow fails when any single provider activity fails

[jorgecuesta]

Problem

ProviderStatus uses Promise.all to fan out per-provider activities. If any one of the processProviderStatus activities exhausts its retries and rejects, the whole workflow execution is marked Failed, even when the other providers' activities completed successfully.

In practice this happens every minute on mainnet — a few external provider endpoints regularly return HTML or "Bad Gateway" instead of JSON (vido.info, weaversnodes.com, poktpool.com, nodefleet.org, kleomed.es). Result: ~100% of ProviderStatus workflow runs end as Failed despite most provider activities succeeding.

Source

apps/middleman-workflows/src/workflows/ProviderStatus.ts:17

await Promise.all(
  providers.map((provider) => processProviderStatus(provider.identity))
);

Suggested fix

Swap Promise.all for Promise.allSettled so providers fail independently:

await Promise.allSettled(
  providers.map((provider) => processProviderStatus(provider.identity))
);

Workflow ends as Completed; individual provider failures remain visible as ActivityTaskFailed events in the workflow history.