Node service crashing due to connection reset

[bryanchriswhite]

Background

• The subquery node queries the configured fetch rpc endpoint to fetch data as it runs.
• The node container exits when it encounters errors and then relies on it's orchestration environment to restart it.
• Cross-cluster (from sandbox to testnet) requests may be triggering the istio circuit breaker

This is the error occurs frequently and causes the deployment to go into crashloop backoff:

2022-08-16T09:40:52.005Z <fetch> ERROR failed to index block at height 977210  Error: read ECONNRESET

[bryanchriswhite]

This seems to only apply in scenarios where the indexer is running in a different cluster than the fetchd node it is querying.

[bryanchriswhite]

Closing as this is a known phenomenon related to network security mechanisms of fetchd and has limited applicability (i.e. non-production environments).