coin_balance_catchup errors, while block_catchup working fine

[ixje]

In my logs a see a lot of the following errors

{"time":"2024-12-19T15:08:46.866Z","severity":"error","message":"failed to fetch: :etimedout","metadata":{"count":100,"fetcher":"coin_balance_catchup","error_count":100}}

while at the same time block_catchup appears to be working fine.

{"time":"2024-12-19T15:08:54.818Z","severity":"info","message":"Index already caught up.","metadata":{"fetcher":"block_catchup","last_block_number":null,"shrunk":false,"missing_block_count":0,"first_block_number":null}}

Everything points to the same private node. I assume that if block_catchup can conclude we're not missing any blocks that the node is reachable for data queries.

What can cause a etimedout for coin_balance_catchup though?

[vbaranov]

A relatively high batch size for a coin balance fetcher can cause a timeout issue at request. Consider decreasing the default 100 batch size via runtime env variable INDEXER_COIN_BALANCES_BATCH_SIZE.

[ixje]

Thanks for the suggestion. I noticed that last_block_number and first_block_number in the log message were both null and started thinking that that might be the reason it outputs Index already caught up. After some more digging it turned out to be a network reach-ability issue after all. Now last_block_number and first_block_number show numbers instead.