Debug subgraph indexing time

[jennijuju]

https://filecoinproject.slack.com/archives/C0717TGU7V2/p1753693062099249

[rjan90]

Goldsky has put up subgraph monitoring at https://monitor.hem.so/status/filecoin, which will ping one of their team members if either a indexing error happens, or if there is more than 1 block latency on the subgraph.

They also mentioned a potential alternative indexing approach that could serve as a long-term solution:

i do want to mention, there may be an alternative indexing approach here that might be the right long-term solution; mirror.

this works by streaming data into a db you own directly, and manages reorgs a bit differently: we'll push data as soon as it's finalized. if a reorg happens, will will upsert/delete data right in your sink automatically. the latency to the tip of the chain should be VERY low, under a second in most cases

[rjan90]

At this point, I don’t believe there are any immediate action items. The next step is likely to monitor the Subgraph Indexing performance over time. If we consistently observe block latency greater than 1, we may need to consider taking action.

[juliangruber]

If we consistently observe block latency greater than 1, we may need to consider taking action.

This is the case for us:

We have disabled logging for the 2 blocks delay case because it happens too often