David Lutterkort

@YassinEldeeb Do you have the schema for the GraphQL server this generates somewhere? It would help in reviewing these changes.

Closing this since it's not been active for a while. Feel free to reopen when you're ready to work on this again

The code changes look fine, but making this configurable has repercussions for the network; I think @leoyvens is in a better position to make a call on how to deal...

Is the main motivation for this to reduce the amount of redundantly stored data, i.e., just one `Token` id vs other attributes like `symbol` etc. ? I wonder if we...

> I wasn't thinking of nesting `tokenStats` within `token` for queries, though it's an interesting idea. My motivation is that it makes the timeseries directly queryable as any other entity,...

> This actually has interesting consequences for the write side. If the query side can handle partial buckets, it probably wouldn't be a big stretch for it to handle missing...

For the current bucket, we'll have to do the rollup at query time. My hope is that by making the default to not include the current bucket, it will only...

Just to log some progress, here's what remains to be done: - [x] Review/testing of PR #5184 - [x] Change how first/last aggregates are computed - define `arg_min`/`arg_max` functions in...

When graph-node starts up, it assumes all subgraphs are not synced; once it processes a block, it determines whether the subgraph is synced and then turns off recording the metric....

I haven't looked at it in detail but supporting IPv6 in addition to IPv4 makes absolute sense and we should support it.